Podcasts about generativeai

Send us a textAgentic AI is rapidly transforming professional services, yet many consultancies remain cautious about its implementation. In this groundbreaking conversation, Professor Joe O'Mahony and Rob Price explore what happens when multiple AI agents work together as a team to tackle complex business challenges.Rob explains how agentic AI differs from traditional AI applications by creating virtual team members with specific expertise, memory capabilities, and tools. These aren't simple chatbots – they're sophisticated systems that can handle complex workflows while maintaining human oversight. The most fascinating examples include agent teams already working in regulated environments for apprenticeship coaching and due diligence processes, reducing tasks from hours to minutes.The conversation takes a profound turn when examining how this technology threatens the traditional consulting pyramid. As AI increasingly handles work previously done by junior consultants, firms face difficult questions about talent development, pricing models, and the very nature of their business. "It's not AI that will steal your job—it's the person using AI who will," warns Rob, highlighting the competitive advantage for those who master these technologies.For young consultants, this shift demands a hybrid skillset combining technical understanding with enhanced human capabilities. The most valuable consultants will be those who build trust, understand context, and create innovative solutions that AI alone cannot. Meanwhile, established professionals might explore what the hosts call "influencer consultancy" – using AI to amplify their personal brand and impact beyond traditional firm structures.The conversation concludes with a balanced view of consulting's future: short-term opportunities for those who embrace AI capabilities, alongside longer-term questions about how value and profits will be distributed. Whether you're leading a consultancy or considering a career in professional services, this episode provides crucial insights into navigating the agentic AI revolution that's already underway.Prof. Joe O'Mahoney helps boutique consultancies scale and exit. Joe's research, writing, speaking and insights can be found at www.joeomahoney.com

The future of learning isn’t static — it’s responsive, reflective, and built for capability. In this follow-up to our first AI episode, Zoe, Michelle, and Ausmed CEO Will Egan explore the next phase of AI in learning: generative learning — and the governance frameworks that make it safe and useful in real-world healthcare settings. From rubrics and role-based prompts to closed systems and guardrails, this episode takes a look into how generative AI can personalise learning, support performance, and finally help us bridge the gap between theory and practice. Contact the show at podcast@ausmed.com.au Follow Ausmed on LinkedIn, Facebook & Instagram Resources: AI and Privacy | Guide AI Bias and Cultural Safety in Aged Care | Guide Using AI in Healthcare Education | Guide How AI and Machine Learning Is Impacting Nursing | Guide What does AI mean for healthcare in Australia? | Guide Why Reflection Matters | Guide Thinking Differently About Change (Pt. 1) | Thought Leadership Smart Strategies for Healthcare Education | Thought Leadership Learn More About Ausmed: https://lnk.bio/ausmed/organisationsSee omnystudio.com/listener for privacy information.

In this episode of "Father and Joe," hosts Father Boniface Hicks and Joe Rockey dive deep into the fast-evolving world of artificial intelligence and its implications on human relationships. With AI becoming an integral part of our lives, it's crucial to understand its impact, especially on the sanctity and quality of human connections. Father Boniface and Joe explore the nuances of trust within relationships in an age where AI can imitate human behavior with uncanny precision. Can technology ever replicate the profound depth of human relationships? Join the hosts as they discuss the dangers of blurring the lines between genuine human interaction and AI-powered communication.The episode looks into the potential pitfalls of AI-generated content and how it may compromise our ability to discern truth from fiction. Our relationship dynamics, whether personal or professional, rely heavily on the trust and authenticity that AI may challenge. This conversation underscores the importance of maintaining robust in-person relationships and developing skills to ensure what we perceive as real is indeed so. Father Boniface touches on the philosophical and theological aspects of these changes, calling listeners to reconsider the value of human connections that transcend mere transactional interactions.Joe brings to light the effects seen in the business and social landscapes, where AI is often used to automate everything from advertising to customer interactions. The hosts discuss the potential saturation and diminishing quality of AI-generated content, which could cause a decline in meaningful human engagement.As Joe and Father Boniface navigate these complex ideas, they challenge listeners to enhance their "relationship muscles" and prioritize cultivating genuine human connections. Whether it's strengthening existing bonds or repairing broken ones, they highlight the critical need for human interaction in our technology-driven world.Tags: Artificial Intelligence, AI Impact, Human Relationships, Trust, Technology and Humanity, Father and Joe, Podcast, Spiritual Guidance, Relationship Skills, AI Concerns, Authentic Connections, Digital Age, AI Content, Communication, Human Interaction, Father Boniface Hicks, Joe Rockey, Personal Development, Spirituality, Theology, AI Challenges, AI Future, Business Impacts, Social Media, Online Interactions, Human Connection, Life Skills, Technological Growth, AI Ethics, Digital Communication, AI Algorithms, Relationship Dynamics, Trust in Technology, Spiritual Reflection, Real vs Fake, New Technology, Human Creativity, Generative AI, AI in Society, Faith and TechnologyHashtags: #ArtificialIntelligence, #AIImpact, #HumanRelationships, #TrustIssues, #TechAndHumanity, #FatherAndJoe, #PodcastTalk, #SpiritualGuidance, #RelationshipSkills, #AIFears, #DigitalConnections, #AIContent, #TrueCommunication, #HumanInteraction, #FatherBonifaceHicks, #JoeRockey, #PersonalGrowth, #Spirituality, #TheologicalTalk, #AIChallenges, #FutureTech, #BusinessImpact, #SocialMedia, #OnlineInteractions, #RealHumanConnection, #LifeSkills, #TechGrowth, #AIethics, #DigitalCommunication, #AIGeneration, #RelationshipDynamics, #TechTrust, #SpiritualReflection, #RealVsFake, #NewTech, #HumanCreativity, #GenerativeAI, #AISociety, #FaithAndTechThis line is here to correct the site's formatting error.

AI is reshaping industries at a rapid pace, but as its influence grows, so do the ethical concerns that come with it.   This episode examines how AI is being applied across sectors such as healthcare, finance, and retail, while also exploring the crucial issue of ensuring that these technologies align with human values.   In this conversation, Lois Houston and Nikita Abraham are joined by Hemant Gahankari, Senior Principal OCI Instructor, who emphasizes the importance of fairness, inclusivity, transparency, and accountability in AI systems.   AI for You: https://mylearn.oracle.com/ou/course/ai-for-you/152601/   Oracle University Learning Community: https://education.oracle.com/ou-community   LinkedIn: https://www.linkedin.com/showcase/oracle-university/   X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.   ---------------------------------------------------- Episode Transcript:   00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:25 Lois: Welcome to the Oracle University Podcast! I'm Lois Houston, Director of Innovation Programs with Oracle University, and with me is Nikita Abraham, Team Lead: Editorial Services. Nikita: Hey everyone! In our last episode, we spoke about how Oracle integrates AI capabilities into its Fusion Applications to enhance business workflows, and we focused on Predictive, Generative, and Agentic AI. Lois: Today, we'll discuss the various applications of AI. This is the final episode in our AI series, and before we close, we'll also touch upon ethical and responsible AI.  01:01 Nikita: Taking us through all of this is Senior Principal OCI Instructor Hemant Gahankari. Hi Hemant! AI is pretty much everywhere today. So, can you explain how it is being used in industries like retail, hospitality, health care, and so on?  Hemant: AI isn't just for sci-fi movies anymore. It's helping doctors spot diseases earlier and even discover new drugs faster. Imagine an AI that can look at an X-ray and say, hey, there is something sketchy here before a human even notices. Wild, right? Banks and fintech companies are all over AI. Fraud detection. AI has got it covered. Those robo advisors managing your investments? That's AI too. Ever noticed how e-commerce companies always seem to know what you want? That's AI studying your habits and nudging you towards that next purchase or binge watch. Factories are getting smarter. AI predicts when machines will fail so they can fix them before everything grinds to a halt. Less downtime, more efficiency. Everyone wins. Farming has gone high tech. Drones and AI analyze crops, optimize water use, and even help with harvesting. Self-driving cars get all the hype, but even your everyday GPS uses AI to dodge traffic jams. And if AI can save me from sitting in bumper-to-bumper traffic, I'm all for it. 02:40 Nikita: Agreed! Thanks for that overview, but let's get into specific scenarios within each industry.  Hemant: Let us take a scenario in the retail industry-- a retail clothing line with dozens of brick-and-mortar stores. Maintaining proper inventory levels in stores and regional warehouses is critical for retailers. In this low-margin business, being out of a popular product is especially challenging during sales and promotions. Managers want to delight shoppers and increase sales but without overbuying. That's where AI steps in. The retailer has multiple information sources, ranging from point-of-sale terminals to warehouse inventory systems. This data can be used to train a forecasting model that can make predictions, such as demand increase due to a holiday or planned marketing promotion, and determine the time required to acquire and distribute the extra inventory. Most ERP-based forecasting systems can produce sophisticated reports. A generative AI report writer goes further, creating custom plain-language summaries of these reports tailored for each store, instructing managers about how to maximize sales of well-stocked items while mitigating possible shortages. 04:11 Lois: Ok. How is AI being used in the hospitality sector, Hemant? Hemant: Let us take an example of a hotel chain that depends on positive ratings on social media and review websites. One common challenge they face is keeping track of online reviews, leading to missed opportunities to engage unhappy customers complaining on social media. Hotel managers don't know what's being said fast enough to address problems in real-time. Here, AI can be used to create a large data set from the tens of thousands of previously published online reviews. A textual language AI system can perform a sentiment analysis across the data to determine a baseline that can be periodically re-evaluated to spot trends. Data scientists could also build a model that correlates these textual messages and their sentiments against specific hotel locations and other factors, such as weather. Generative AI can extract valuable suggestions and insights from both positive and negative comments. 05:27 Nikita: That's great. And what about Financial Services? I know banks use AI quite often to detect fraud. Hemant: Unfortunately, fraud can creep into any part of a bank's retail operations. Fraud can happen with online transactions, from a phone or browser, and offsite ATMs too. Without trust, banks won't have customers or shareholders. Excessive fraud and delays in detecting it can violate financial industry regulations. Fraud detection combines AI technologies, such as computer vision to interpret scanned documents, document verification to authenticate IDs like driver's licenses, and machine learning to analyze patterns. These tools work together to assess the risk of fraud in each transaction within seconds. When the system detects a high risk, it triggers automated responses, such as placing holds on withdrawals or requesting additional identification from customers, to prevent fraudulent activity and protect both the business and its client. 06:42 Nikita: Wow, interesting. And how is AI being used in the health industry, especially when it comes to improving patient care? Hemant: Medical appointments can be frustrating for everyone involved—patients, receptionists, nurses, and physicians. There are many time-consuming steps, including scheduling, checking in, interactions with the doctors, checking out, and follow-ups. AI can fix this problem through electronic health records to analyze lab results, paper forms, scans, and structured data, summarizing insights for doctors with the latest research and patient history. This helps practice reduced costs, boost earnings, and deliver faster, more personalized care. 07:32 Lois: Let's take a look at one more industry. How is manufacturing using AI? Hemant: A factory that makes metal parts and other products use both visual inspections and electronic means to monitor product quality. A part that fails to meet the requirements may be reworked or repurposed, or it may need to be scrapped. The factory seeks to maximize profits and throughput by shipping as much good material as possible, while minimizing waste by detecting and handling defects early. The way AI can help here is with the quality assurance process, which creates X-ray images. This data can be interpreted by computer vision, which can learn to identify cracks and other weak spots, after being trained on a large data set. In addition, problematic or ambiguous data can be highlighted for human inspectors. 08:36 Oracle University's Race to Certification 2025 is your ticket to free training and certification in today's hottest tech. Whether you're starting with Artificial Intelligence, Oracle Cloud Infrastructure, Multicloud, or Oracle Data Platform, this challenge covers it all! Learn more about your chance to win prizes and see your name on the Leaderboard by visiting education.oracle.com/race-to-certification-2025. That's education.oracle.com/race-to-certification-2025. 09:20 Nikita: Welcome back! AI can be used effectively to automate a variety of tasks to improve productivity, efficiency, cost savings. But I'm sure AI has its constraints too, right? Can you talk about what happens if AI isn't able to echo human ethics?  Hemant: AI can fail due to lack of ethics.  AI can spot patterns, not make moral calls. It doesn't feel guilt, understand context, or take responsibility. That is still up to us.  Decisions are only as good as the data behind them. For example, health care AI underdiagnosing women because research data was mostly male. Artificial narrow intelligence tends to automate discrimination at scale. Recruiting AI downgraded resumes just because it had a word "women's" (for example, women's chess club). Who is responsible when AI fails? For example, if a self-driving car hits someone, we cannot blame the car. Then who owns the failure? The programmer? The CEO? Can we really trust corporations or governments having programmed the use of AI not to be evil correctly? So, it's clear that AI needs oversight to function smoothly. 10:48 Lois: So, Hemant, how can we design AI in ways that respect and reflect human values? Hemant: Think of ethics like a tree. It needs all parts working together. Roots represent intent. That is our values and principles. The trunk stands for safeguards, our systems, and structures. And the branches are the outcomes we aim for. If the roots are shallow, the tree falls. If the trunk is weak, damage seeps through. The health of roots and trunk shapes the strength of our ethical outcomes. Fairness means nothing without ethical intent behind it. For example, a bank promotes its loan algorithm as fair. But it uses zip codes in decision-making, effectively penalizing people based on race. That's not fairness. That's harm disguised as data. Inclusivity depends on the intent sustainability. Inclusive design isn't just a check box. It needs a long-term commitment. For example, controllers for gamers with disabilities are only possible because of sustained R&D and intentional design choices. Without investment in inclusion, accessibility is left behind. Transparency depends on the safeguard robustness. Transparency is only useful if the system is secure and resilient. For example, a medical AI may be explainable, but if it is vulnerable to hacking, transparency won't matter. Accountability depends on the safeguard privacy and traceability. You can't hold people accountable if there is no trail to follow. For example, after a fatal self-driving car crash, deleted system logs meant no one could be held responsible. Without auditability, accountability collapses. So remember, outcomes are what we see, but they rely on intent to guide priorities and safeguards to support execution. That's why humans must have a final say. AI has no grasp of ethics, but we do. 13:16 Nikita: So, what you're saying is ethical intent and robust AI safeguards need to go hand in hand if we are to truly leverage AI we can trust. Hemant: When it comes to AI, preventing harm is a must. Take self-driving cars, for example. Keeping pedestrians safe is absolutely critical, which means the technology has to be rock solid and reliable. At the same time, fairness and inclusivity can't be overlooked. If an AI system used for hiring learns from biased past data, say, mostly male candidates being hired, it can end up repeating those biases, shutting out qualified candidates unfairly. Transparency and accountability go hand in hand. Imagine a loan rejection if the AI's decision isn't clear or explainable. It becomes impossible for someone to challenge or understand why they were turned down. And of course, robustness supports fairness too. Loan approval systems need strong security to prevent attacks that could manipulate decisions and undermine trust.  We must build AI that reflects human values and has safeguards. This makes sure that AI is fair, inclusive, transparent, and accountable.  14:44 Lois: Before we wrap, can you talk about why AI can fail? Let's continue with your analogy of the tree. Can you explain how AI failures occur and how we can address them? Hemant: Root elements like do not harm and sustainability are fundamental to ethical AI development. When these roots fail, the consequences can be serious. For example, a clear failure of do not harm is AI-powered surveillance tools misused by authoritarian regimes. This happens because there were no ethical constraints guiding how the technology was deployed. The solution is clear-- implement strong ethical use policies and conduct human rights impact assessment to prevent such misuse. On the sustainability front, training AI models can consume massive amount of energy. This failure occurs because environmental costs are not considered. To fix this, organizations are adopting carbon-aware computing practices to minimize AI's environmental footprint. By addressing these root failures, we can ensure AI is developed and used responsibly with respect for human rights and the planet. An example of a robustness failure can be a chatbot hallucinating nonexistent legal precedence used in court filings. This could be due to training on unverified internet data and no fact-checking layer. This can be fixed by grounding in authoritative databases. An example of a privacy failure can be AI facial recognition database created without user consent. The reason being no consent was taken for data collection. This can be fixed by adopting privacy-preserving techniques. An example of a fairness failure can be generated images of CEOs as white men and nurses as women, minorities. The reason being training on imbalanced internet images reflecting societal stereotypes. And the fix is to use diverse set of images. 17:18 Lois: I think this would be incomplete if we don't talk about inclusivity, transparency, and accountability failures. How can they be addressed, Hemant? Hemant: An example of an inclusivity failure can be a voice assistant not understanding accents. The reason being training data lacked diversity. And the fix is to use inclusive data. An example of a transparency and accountability failure can be teachers could not challenge AI-generated performance scores due to opaque calculations. The reason being no explainability tools are used. The fix being high-impact AI needs human review pathways and explainability built in. 18:04 Lois: Thank you, Hemant, for a fantastic conversation. We got some great insights into responsible and ethical AI. Nikita: Thank you, Hemant! If you're interested in learning more about the topics we discussed today, head over to mylearn.oracle.com and search for the AI for You course. Until next time, this is Nikita Abraham…. Lois: And Lois Houston, signing off! 18:26 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.  

Security leaders from CyberArk, Fortra, and Sysdig share actionable strategies for securely implementing generative AI and reveal real-world insights on data protection and agent management.Topics Include:Panel explores practical security approaches for GenAI from prototype to productionThree-phase framework discussed: planning, pre-production, and production security considerationsSecurity must be built-in from start - data foundation is criticalUnderstanding data location, usage, transformation, and regulatory requirements is essentialFortra's security conglomerate approach integrates with AWS native tools and partnersMachine data initially easier for compliance - no PII or HIPAA concernsIdentity paradigm shift: agents can dynamically take human and non-human roles97% of organizations using AI tools lack identity and access policiesSecurity responsibility increases as you move up the customization stackOWASP Top 10 for GenAI addresses prompt injection and data poisoningRigorous model testing including adversarial attacks before deployment is crucialSysdig spent 6-9 months stress testing their agent before production releaseTension exists between moving fast and implementing proper security controlsDifferent security approaches needed based on data sensitivity and model usageZero-standing privilege and intent-based policies critical for agent managementMulti-agent systems create "Internet of Agents" with exponentially multiplying risksDiscovery challenge: finding where GenAI is running across enterprise environmentsAPI security and gateway protection becoming critical with acceptable latencyTop customer need: translating written AI policies into actionable controlsThreat modeling should focus on impact rather than just vulnerability severityParticipants:Prashant Tyagi - Go-To-Market Identity Security Technology Strategy Lead, CyberArkMike Reed – Field CISO, Cloud Security & AI, FortraZaher Hulays – Vice President Strategic Partnerships, SysdigMatthew Girdharry - WW Leader for Observability & Security Partnerships, Amazon Web ServicesFurther Links:CyberArk: Website – LinkedIn – AWS MarketplaceFortra: Website – LinkedIn – AWS MarketplaceSysdig: Website – LinkedIn – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Can your AI systems be tricked into leaking data? Learn how red teaming can expose hidden vulnerabilities and what you can do to build better defenses.

Brian Mendenhall, Worldwide Head, Security & Identity Partner Specialists of Amazon Web Services, reveals the insider framework for transforming enterprise AI security, including the three-pillar approach and partnership strategies that leading companies use to navigate AI governance challenges.Topics Include:At AWS everything starts with security as core principleConsulting partners follow three-phase model: assess, remediate, then fully manage securityTraditional security framework covers threat detection, incident response, and data protectionAI compliance spans multiple governance bodies with stacking requirements and regulationsEU AI Act affects any company globally if Europeans access their applicationsThree pillars: security OF AI, AI FOR security, security FROM AI attacksAWS launches AI security competency program with specialized partner categories and certificationsEnterprise AI spans five risk levels from consumer apps to self-trained modelsLegal liability dramatically increases as you move toward custom AI implementationsSafety means preventing harm; security means preventing breaches - both critical distinctionsCurrent AI hallucination rates hit 65-75% across major platforms like PalantirShared responsibility model determines who's liable when AI security tools failIndustry evolution progresses from machine learning to generative AI to autonomous agentsMajor prototype-to-production gap caused by governance, security, and scalability challengesSuccessful AWS partnerships require clear use cases, differentiation, and targeted go-to-market strategyParticipants:Brian Mendenhall - WW Head, Security & Identity Partner Specialists, Amazon Web ServicesSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Want to make AI work for your business? In today's episode, Lois Houston and Nikita Abraham continue their discussion of AI in Oracle Fusion Applications by focusing on three key AI capabilities: predictive, generative, and agentic.   Joining them is Principal Instructor Yunus Mohammed, who explains how predictive, generative, and agentic AI can optimize efficiency, support decision-making, and automate tasks—all without requiring technical expertise.   AI for You: https://mylearn.oracle.com/ou/course/ai-for-you/152601/   Oracle University Learning Community: https://education.oracle.com/ou-community   LinkedIn: https://www.linkedin.com/showcase/oracle-university/   X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.   ------------------------------------------------------------   Episode Transcript: 00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:25 Nikita: Welcome to the Oracle University Podcast! I'm Nikita Abraham, Team Lead: Editorial Services with Oracle University, and with me is Lois Houston, Director of Innovation Programs. Lois: Hi there! In our last episode, we explored the essential components of the Oracle AI stack and spoke about Oracle's suite of AI services.  Nikita: Yeah, and in today's episode, we're going to go down a similar path and take a closer look at the AI functionalities within Oracle Fusion Applications. 00:53 Lois: With us today is Principal Instructor Yunus Mohammed. Hi Yunus! It's lovely to have you back with us. For anyone who doesn't already know, what are Oracle Fusion Cloud Applications?  Yunus: Oracle Fusion Applications are a suite of cloud-based enterprise applications designed to run for your business across finance, HR, supply chain, sales, services and more, all on a unified platform. They are designed to help enterprises operate smarter, faster by embedding AI directly into business process. That means better forecasts in finance, faster hiring decisions in HR, and optimized supply chains, and more personalized customer experience.  01:42 Nikita: And we know they've been built for today's fast-paced, AI-driven business environment. So, what are the different functional pillars within Oracle Fusion Apps? Yunus: The first one is the ERP, Enterprise Resource Planning, which supports financials, procurements, and project management. It's the backbone of many organizations, or day-to-day operations. HCM or Human Capital Management, handles workforce-related processes such as hiring, payroll, performance, and talent development, helping HR teams operate more efficiently. SCM, the Supply Chain Management, enables businesses to manage their logistics, inventory, and suppliers and manufacturers in the business. It's particularly critical in industries with complex operations like retail and manufacturing. The CX, which is the Customer Experience, covers the full customer life cycle, which includes sales, marketing, and service. These models help the businesses connect with their customers more personally and proactively, whether through the targeted campaigns or responsive support.  03:02 Lois: Yunus, what sets Fusion apart? Yunus: What sets Fusion apart is how these applications work seamlessly together. They share data natively and continuously improve with AI and automation, giving you not just tools, but intelligence at scale.  Oracle applications are built to be AI first, with a complete suite of finance, supply chain, manufacturing, HR, sales, service, and marketing, that is tightly coupled with our industry and data intelligence applications. The easiest and the most effective way to start building your organization's AI muscle is with AI embedded in Fusion applications. For example, if the customer needs to return a defective product, the service representative simply clicks on Ask Oracle for the answers. Since the AI agent is embedded in the application, it has contextual information about the customer, the order, and any special service, contract, or any other feature that is required for this process. The AI agent automatically figures out the return policy, including the options to send a replacement product immediately or offer a discount for the inconvenience, and also expedite shipping. Another AI agent sends a personalized email confirming details of the return, and different AI agent creates the replacement order for fulfillment and shipping. Our AI-embedded Fusion Applications can automate an end-to-end business process from service request to return order to fulfillment and shipping and then accounting.  These are pre-built and tested so that all the worry and hard work is removed from the implementation point of view. They cover the core workflows. Basically, they address tasks that form part of the organization's core workflow. User requires no technical knowledge in the scenarios.  05:16 Lois: That's great! So, you don't need to be an AI expert or a data scientist to get going. Yunus: The outcomes are super fast in business softwares and context is everything. Just having the right information isn't enough. This is about having the information in the right place at the right time for it to be instantly actionable. They are ready from day one and can be optimized over time. They are powerful out of the box and only get better with day-to-day processes and performance. 05:55 Are you working towards an Oracle Certification this year? Join us at one of our certification prep live events in the Oracle University Learning Community. Get insider tips from seasoned experts and learn from others who have already taken their certifications. Go to community.oracle.com/ou to jump-start your journey towards certification today!  06:20 Nikita: Welcome back! So, when we talk about the AI capabilities in Fusion apps, I know we have different types. Can you tell us more about them?  Yunus: Predictive AI is where it all started. These models analyze historical patterns and data to anticipate what might happen next. For example, predicting employee attrition, forecasting demand in supply chain, or flagging potential late payments in finance workflows. These are embedded into business processes to surface insights before action is needed. Then we have got the generative AI, which takes this a step more further. Instead of just providing insights, it creates content, such as auto-generating job descriptions, summarizing performance reviews, or even crafting draft responses to supplier queries. This saves time and boosts productivity across functions like HR, CX, and procurement. Last but not the least, we have got the agentic AI, which is the most advanced layer. These agents don't just provide suggestions, they take actions on behalf of the users. Think of an agent that not only recommends actions in a workflow, but also executes them, creating tasks, filling tickets, updating systems, and communicating with stakeholders, all autonomously but under user control. And importantly, many business scenarios today benefit from a blend of these types. For example, an AI assistant in Fusion HCM might predict employees turnover, which is predictive AI, generates tailored retention plans, which is generative, and it is generative AI, and initiate outreach or next steps, which is done by the process of agents, which is called agentic AI. So, Oracle integrates these capabilities in a harmonious way, enabling users to act faster, personalize at scale, and drive better business outcomes.  08:39 Lois: Ok, let's get into the specifics. How does Oracle use predictive AI across its Fusion apps, helping businesses anticipate what's coming and act proactively.  Yunus: So in HCM, things like recommended jobs, in this, candidates visiting a potential employer's website encountered an improved online experience, whereby if they have uploaded their resumes, they will be shown job opportunities that match their skills and experience mix. This helps candidates who are unsure what to search by showing them roles and titles they may not have considered. Time to hire provides an estimated as to how long it will take for an HR team to fill an open role, but this is really useful not only in terms of planning, recruitment, but also in terms of understanding whether you might need some temporary cover and for how long will it actually take the process to complete. In the process of supply chain management, the predictive AI is leveraged to revolutionize transit time and estimated time of arrival, which is called as the predictive analysis, enhancing efficiency, and optimizing operations. It can flag abnormal patterns in supply or inventory. For example, if a batch of parts is behaving differently in the production line and predict future demands, helping avoid overstocking or stockouts is a process that can be done by using the SCM predictive analysis or predictive AI. In ERPs, where you can audit your expenses, plan for future expenses, and do dynamic discounting for vendors who are likely to accept earlier payments or earlier payment discounts, it can also speed up reimbursements by automated expense entries. In CX, you have the options to go with adaptive intelligence for sales, which helps representatives prioritize the leads and the likelihood that a specific lead will close, helping representatives focus their time and effort. So predictive scheduling and routing in service delivery ensures that the right resource is assigned to the right customer at the right time, boosting operational efficiency and customer satisfaction, also known as fatigue analysis. 11:23 Lois: Now let's shift our focus to generative AI. How does Oracle implement generative AI across HCM, ERP, Supply Chain, and CX? Yunus: So, in HCM, the generative AI can automatically generate performance review summaries from raw data, saving time for HR teams, and can help you in providing candidates with summaries of their interview process, feedback, and next steps, all auto generated. With AI assistance, goal creation for employees can be automated, and the system analyzes performance data and trends to propose meaningful and attainable goals, aligning them with organizational objectives and employee capabilities. In SCM, similarly, the generative AI process helps you in defining drafting summaries of purchase orders. So generative AI can automatically create clear, readable synopses, and can be summarized with complex negotiations and discussions, making it easier for supply chain managers to analyze supplier proposals, track negotiations, processes, and understand key takeaways. With predictive AI embedded, it is helping you to leverage to help generate the repairs of master definitions of summaries, and can generate descriptions for item based on their specification, helping product teams automatically generate catalog contents. With ERPs, you can automate the creation of business reports, offering more insights and actionable narratives, rather than just showing the raw data. The AI can provide context, interpretations, and recommendations. AI can also take raw project data and generate a comprehensive, easy-to-read project status, reports that stakeholders can quickly review. In CX, we have got service request summarization, which can provide these long summaries for the customer services and the tickets that have been requested by the customers, allowing support teams to understand the key points in the fraction of time, and can also create knowledge base articles directly from common service requests or inquiries, which not only improves internal knowledge management but also empowers customers by enabling self-service. So generative AI can automatically generate success stories or case studies from successful opportunities or sales, which can be used as marketing content or for internal knowledge sharing. 14:20 Nikita: And what about Oracle's Agentic AI? What are its capabilities across the different pillars? Yunus: In HCM, Agentic AI handles the end-to-end onboarding experience, from explaining policies to guiding document submissions, even booking orientation sessions, allowing the HR staff to focus on human engagement. It can further support HR teams during performance review cycles by surfacing high potential employees, pulling in performance data, and recommending next actions like promotions or learning paths. It helps manage time with requests by checking eligibility, policy constraints, and suggesting appropriate substitutes, reducing administrative frictions and errors. With SCM, the Agentic AI Fusion Applications act as a real time Assistant to ensure buyers follow procurement policies, and reducing compliance risk and manual errors. It can also support sales representatives with real-time insights and next best actions during the quoting or ordering process, improving customer satisfaction and sales performance. With ERP, you can handle document intake, extraction, and routing, saving significant time on manual document management across financial functions using Fusion Applications. AI automates reconciliation tasks by matching transactions, flagging anomalies, and suggesting resolutions. It helps you in reducing close cycle timelines and continuously analyzes profit margins. And it recommends the pricing adjustments that can be done in your ERPs. In CX, the Agentic AI Fusion Application supports staff by instantly compiling full customer histories, orders, service requests, interactions, and can act like a real-time assistant, summarizing open tickets and resolutions, helping agents take over or escalate without needing to dig through the notes, and dynamically adjust technicals and technician routes based on traffic, priority, or cancelation, increasing the field efficiency and customer satisfaction. 17:04 Lois: Thank you so much, Yunus. To learn more about the topics covered today, visit mylearn.oracle.com and search for the AI for You course. Nikita: Join us next week as we cover how AI is being applied across sectors like healthcare, finance, and retail, and tackle the big question: how do we keep these technologies aligned with human values? Until then, this is Nikita Abraham… Lois: And Lois Houston, signing off! 17:30 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.

Caitlin Anderson, Intel's Americas Sales GM shares which industries are leading AI adoption, where the biggest untapped opportunities lie, and why AI spending is expected to double by 2028. With special guest Piyush Sharrma of Tuskira.aiTopics Include:Caitlin Anderson discusses Intel-AWS partnership and generative AI trends accelerating businessIntel's AI journey spans decades: analytics since 1980s, natural language processing 2000sComputer vision remains major use case from edge computing to data centersGenerative AI and AI agents are the latest wave, with agents collaborating togetherIntel uses AI internally for manufacturing automation in highly sensitive fab environmentsRobotics and AI optimize quality control, system monitoring, and technician productivityAI spending growth spans all industries, with significant acceleration expected through 2028Software services, healthcare, and financial services lead current AI adoption and experimentationEducation, government, retail, and energy represent major untapped growth opportunities aheadIntel-AWS partnership spans 20 years, featuring custom silicon and broad CPU portfolioTuskira CEO Piyush Sharrma explains cybersecurity "perfect storm" where attackers weaponize same AI toolsSuccess requires ecosystem partnerships - no single company can solve complex AI challengesParticipants:Caitlin Anderson - Corporate Vice President, GM Americas Sales, IntelPiyush Sharrma – CEO and Co-Founder, Tuskira.aiSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Aditya Vasudevan, Cohesity's cyber recovery expert, shares battle-tested insights from defending Fortune 100 companies against AI-powered cyberattacks.Topics Include:Cohesity protects 85% of Fortune 100 data with battle-tested cyber recovery experienceTop 10 cyber adversaries target organizations; Cohesity has defended against most major threatsGenAI adopted by 100 million users in two months, creating unprecedented security challengesNew AI threats include prompt injection, synthetic identities, shadow AI, and supply vulnerabilitiesAttackers now use AI for sophisticated phishing, automated malware, and accelerated attack chainsReal companies completely banned AI after code leaks, misuse incidents, and data concernsThree-pillar security approach: fight AI with AI, enhanced training, and automated workflowsSecure AI design requires private deployments, complete traceability, and role-based access controlsAmazon Bedrock offers built-in guardrails, private VPCs, and enterprise monitoring capabilitiesCohesity's Gaia demonstrates secure AI with RAG architecture and permission-aware data accessResilience strategy combines immutable backups, anomaly detection, and recovery automation for incidentsProper AI security reduces cyber insurance premiums and prevents costly downtime disastersParticipants:Aditya Vasudevan - GVP of Cyber Resiliency, Cohesity Further Links:Cohesity: Website | LinkedIn | AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Tyler Warden, SVP of Product at Sonatype, shares surprising research on security, productivity and prioritization, with actionable strategies for organizational transformation. Topics Include:Tyler from Sonatype (Maven creators) shares research on security culture in developmentSecurity is more cultural than tooling, with rising supply chain attacksDevelopment speeds up while global regulations rapidly change across marketsTyler's background: wanted to be a Broadway conductor, not tech speakerBeethoven's 9th Symphony story: nephew missed a dot, changing tempo foreverWe can "be the dot" - small changes creating big organizational impactThree organization types: Leaders (collaborative), Adapters (balanced), Protectors (security-first)Leaders achieve best productivity and security but face executive skepticismResearch reveals balanced teams outperform purely security-focused or productivity-focused approachesHigh-performance teams go faster AND stay more secure than alternatives"Yes" philosophy from improv comedy: fun happens when we enable innovationApply proven supply chain principles from manufacturing to software development security Participants:Tyler Warden – Senior Vice President, Product, SonatypeFurther Links:Sonatype: Website | LinkedIn | AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Organizations across various industries are increasingly seeking to harness generative AI's capabilities to enhance productivity and operational efficiency. But while generative AI technologies can provide organizations with multiple opportunities, there are also risks unless integration is carried out with caution. In this episode of Risk in Context, Marsh's Gregory Eskins and Mercer's Adriana O'Kain discuss the multiple opportunities that generative AI presents, look at how the use of these technologies is evolving, and provide actions that senior leaders should consider to address technical, process, and people implications. You can access a transcript of the episode here. Read our series on debunking AI-generated myths. For more insights and insurance and risk management solutions, follow Marsh on LinkedIn and X and visit marsh.com.

In this episode, Lois Houston and Nikita Abraham are joined by Principal Instructor Yunus Mohammed to explore Oracle's approach to enterprise AI. The conversation covers the essential components of the Oracle AI stack and how each part, from the foundational infrastructure to business-specific applications, can be leveraged to support AI-driven initiatives.   They also delve into Oracle's suite of AI services, including generative AI, language processing, and image recognition.     AI for You: https://mylearn.oracle.com/ou/course/ai-for-you/152601/   Oracle University Learning Community: https://education.oracle.com/ou-community   LinkedIn: https://www.linkedin.com/showcase/oracle-university/   X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.   -------------------------------------------------------------   Episode Transcript:  00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:25 Lois: Welcome to the Oracle University Podcast! I'm Lois Houston, Director of Innovation Programs with Oracle University, and with me is Nikita Abraham, Team Lead: Editorial Services. Nikita: Hey everyone! In our last episode, we discussed why the decision to buy or build matters in the world of AI deployment. Lois: That's right, Niki. Today is all about the Oracle AI stack and how it empowers not just developers and data scientists, but everyday business users as well. Then we'll spend some time exploring Oracle AI services in detail.  01:00 Nikita: Yunus Mohammed, our Principal Instructor, is back with us today. Hi Yunus! Can you talk about the different layers in Oracle's end-to-end AI approach? Yunus: The first base layer is the foundation of AI infrastructure, the powerful compute and storage layer that enables scalable model training and inferences. Sitting above the infrastructure, we have got the data platform. This is where data is stored, cleaned, and managed. Without a reliable data foundation, AI simply can't perform. So base of AI is the data, and the reliable data gives more support to the AI to perform its job. Then, we have AI and ML services. These provide ready-to-use tools for building, training, and deploying custom machine learning models. Next, to the AI/ML services, we have got generative AI services. This is where Oracle enables advanced language models and agentic AI tools that can generate content, summarize documents, or assist users through chat interfaces. Then, we have the top layer, which is called as the applications, things like Fusion applications or industry specific solutions where AI is embedded directly into business workflows for recommendations, forecasting or customer support. Finally, Oracle integrates with a growing ecosystem of AI partners, allowing organizations to extend and enhance their AI capabilities even further. In short, Oracle doesn't just offer AI as a feature. It delivers it as a full stack capability from infrastructure to the layer of applications. 02:59 Nikita: Ok, I want to get into the core AI services offered by Oracle Cloud Infrastructure. But before we get into the finer details, broadly speaking, how do these services help businesses? Yunus: These services make AI accessible, secure, and scalable, enabling businesses to embed intelligence into workflows, improve efficiency, and reduce human effort in repetitive or data-heavy tasks. And the best part is, Oracle makes it easy to consume these through application interfaces, APIs, software development kits like SDKs, and integration with Fusion Applications. So, you can add AI where it matters without needing a data scientist team to do that work.  03:52 Lois: So, let's get down to it. The first core service is Oracle's Generative AI service. What can you tell us about it?  Yunus: This is a fully managed service that allows businesses to tap into the power of large language models. You can actually work with these models from scratch to a well-defined develop model. You can use these models for a wide range of use cases like summarizing text, generating content, answering questions, or building AI-powered chat interfaces.  04:27 Lois: So, what will I find on the OCI Generative AI Console? Yunus: OCI Generative AI Console highlights three key components. The first one is the dedicated AI cluster. These are GPU powered environments used to fine tune and host your own custom models. It gives you control and performance at scale. Then, the second point is the custom models. You can take a base language model and fine tune it using your own data, for example, company manuals or HR policies or customer interactions, which are your own personal data. You can use this to create a model that speaks your business language. And last but not the least, the endpoints. These are the interfaces through which your application connect to the model. Once deployed, your app can query the model securely and at different scales, and you don't need to be a developer to get started. Oracle offers a playground, which is a non-core environment where you can try out models, craft parameters, and test responses interactively. So overall, the generative AI service is designed to make enterprise-grade AI accessible and customizable. So, fitting directly into business processes, whether you are building a smart assistant or you're automating the content generation process.  06:00 Lois: The next key service is OCI Generative AI Agents. Can you tell us more about it?  Yunus: OCI Generative AI agents combines a natural language interface with generative AI models and enterprise data stores to answer questions and take actions. The agent remembers the context, uses previous interactions, and retrieves deeper product speech details. They aren't just static chat bots. They are context aware, grounded in business data, and able to handle multi-turns, follow-up queries with relevant accurate responses, and driving productivity and decision-making across departments like sales, support, or operations. 06:54 Oracle University's Race to Certification 2025 is your ticket to free training and certification in today's hottest tech. Whether you're starting with Artificial Intelligence, Oracle Cloud Infrastructure, Multicloud, or Oracle Data Platform, this challenge covers it all! Learn more about your chance to win prizes and see your name on the Leaderboard by visiting education.oracle.com/race-to-certification-2025. That's education.oracle.com/race-to-certification-2025. 07:37 Nikita: Welcome back! Yunus, let's move on to the OCI Language service.  Yunus: OCI Language helps business understand and process natural language at scale. It uses pretrained models, which means they are already trained on large industry data sets and are ready to be used right away without requiring AI expertise. It detects over 100 languages, including English, Japanese, Spanish, and more. This is great for global business that receive multilingual inputs from customers. It works with identity sentiments. For different aspects of the sentence, for example, in a review like, “The food was great, but the service sucked,” OCI Language can tell that food has a positive sentiment while service has a negative one. This is called aspect-based sentiment analysis, and it is more insightful than just labeling the entire text as positive or negative. Then we have got to identify key phrases representing important ideas or subjects. So, it helps in extracting these key phrases, words, or terms that capture the core messages. They help automate tagging, summarizing, or even routing of content like support tickets or emails.  In real life, the businesses are using this for customer feedback analysis, support ticket routing, social media monitoring, and even regulatory compliances.  09:21 Nikita: That's fantastic. And what about the OCI Speech service?  Yunus: The OCI Speech is an AI service that transcribes speech to text. Think of it as an AI-powered transcription engine that listens to the spoken English, whether in audio or video files, and turns it into usable and searchable and readable text. It provides timestamps, so you know exactly when something was said. A valuable feature for reviewing legal discussions, media footages, or compliance audits. OCI Speech even understands different speakers. You don't need to train this from scratch. It is pre-trained model hosted on an API. Just send your audio to the service, and you get an accurate timestamp text back in return. 10:17 Lois: I know we also have a service for object detection… called OCI Vision?  Yunus: OCI Vision uses pretrained, deep learning models to understand and analyze visual content. Just like a human might, you can upload an image or videos, and the AI can tell you what is in it and where they might be useful. There are two primary use cases, which you can use this particular OCI Vision for. One is for object detection. You have got a red color car. So OCI Vision is not just identifying that's a car. It is detecting and labeling parts of the car too, like the bumper, the wheels, the design components. This is a critical in industries like manufacturing, retail, or logistics. For example, in quality control, OCI Vision can scan product images to detect missing or defective parts automatically.  Then we have got the image classification. This is useful in scenarios like automated tagging of photos, managing digital assets, classifying this particular scene or context of this particular scene. So basically, when we talk about OCI Vision, which is actually a fully managed, no complex model training is required for this particular service. It's available via API. It is also working with defining their own custom model for working with the environments. 11:51 Nikita: And the final service is related to text and called OCI Document Understanding, right? Yunus: So OCI Document Understanding allows businesses to automatically extract structured insights from unstructured documents like invoices, contracts, recipes, and also sometimes resumes, or even business documents. 12:13 Nikita: And how does it work? Yunus: OCI reads the content from the scanned document. The OCR is smarter. It recognizes both printed and handwritten text. Then determines what type of document it is. So document classification is done. Text recognition recognizes text, then classifies the document. For example, if this is a purchase order, or bank statement, or any medical report. If your business handles documents in multiple languages, then the AI can actually help in language detection also, which helps you in routing the language or translating that particular language. Many documents contain structured data in table format. Think pricing tables or line items. OCI will help you in extracting these with high accuracy for reporting on feeding into ERP systems. And finally, I would say the key value extraction. It puts our critical business values like invoice numbers, payment amounts, or customer names from fields that may not always allow a fixed format. So, this service reduces the need for manual review, cuts down processes time, and ensures high accuracy for your system. 13:36 Lois: What are the key takeaways our listeners should walk away with after this episode? Yunus: The first one, Oracle doesn't treat AI as just a standalone tool. Instead, AI is integrated from the ground up. Whether you're talking about infrastructure, data platforms, machine learning services, or applications like HCM, ERP, or CX. In real world, the Oracle AI Services prioritize data management, security, and governance, all essential for enterprise AI use cases. So, it is about trust. Can your AI handle sensitive data? Can it comply with regulations? Oracle builds its AI services with strong foundation in data governance, robust security measures, and tight control over data residency and access. So this makes Oracle AI especially well-suited for industries like health care, finance, logistics, and government, where compliance and control aren't optional. They are critical.   14:44 Nikita: Thank you for another great conversation, Yunus. If you're interested in learning more about the topics we discussed today, head on over to mylearn.oracle.com and search for the AI for You course.  Lois: In our next episode, we'll get into Predictive AI, Generative AI, Agentic AI, all with respect to Oracle Fusion Applications. Until then, this is Lois Houston… Nikita: And Nikita Abraham, signing off! 15:10 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.  

Learn how Amazon Bedrock enabled Prophix to build enterprise-grade AI agents that transform secure CFO workflows, while delivering real-time financial intelligence through advanced agentic architecture. Topics Include:AWS and Prophix leaders discuss building autonomous AI agents for financial managementAI agents defined: autonomous systems that reason, plan, and execute tasks independentlyEvolution from simple chatbots to collaborative multi-agent systems solving complex problemsCore agent components: cognitive planning module, memory systems, and external tool integrationsProphix: 30-year financial software company serving 3,500 CFO offices globally across industriesProphix One Intelligence: platform-level AI service powering predictions, analysis, and automationCustomer concerns addressed: data privacy, role-based security, accuracy, and cost controlRejected "models in the sky" approach for AWS Bedrock's managed, controllable infrastructureAgentic architecture: LLMs generate API parameters instead of processing massive datasetsReal-time data access, automatic security inheritance, and A/B testing capabilities achievedLive demo: automated budgeting workflows, natural language queries, and autonomous task executionAWS introduces AgentCore platform to simplify agent development for enterprise customersParticipants:Anurag Yagnik – Chief Technology Officer, ProphixDeborshi Choudhury – Sr Solutions Architect – ISV, Amazon Web ServicesFurther Links:Prophix: Website | LinkedIn See how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Dean Teffer of Arctic Wolf reveals how they process 8 trillion weekly security observations to find "a needle in a stack of needles," and breaks down real-world GenAI lessons learned.Topics Include:Dean Teffer, VP of AI at Arctic Wolf, discusses company's GenAI journeyArctic Wolf: decade-old security operations company serving mid-market customers globallyOperates massive security operation center, now launching AI-powered productsAI agent recently identified Black Basta ransomware attack, enabling rapid containmentDean's 15+ years in cybersecurity: traditional ML focused on detectionGenAI breakthrough allows natural language interaction with security modelsArctic Wolf processes 8 trillion weekly observations, correlating suspicious activitiesChallenge: finding specific threats in "stack of needles," not haystackSuccess measured by making human analysts faster, more consistent, scalableEvolved from treating GenAI like traditional ML to integrated workflowsKey misconception: GenAI isn't magic, needs proper data and reasoningAdvice: start with existing challenges, build flexible systems for adaptationGenAI excels at summarizing information and supporting complex decisionsFuture vision: AI handles routine threats, humans focus on creativityDemocratizing machine learning capabilities to broader range of subject expertsParticipants:Dean Teffer – Vice President of AI, Arctic WolfFurther Links:Arctic Wolf: Website | LinkedIn | AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

CyberArk's technology leader discusses their strategy for securing against AI threats, protecting agentic AI systems, and their vision for the future in an increasingly AI-driven cybersecurity landscape.Topics Include:CyberArk celebrates recent exciting news while discussing their incredible cybersecurity journeyFounded in 1999, CyberArk pioneered privilege access management and expanded into comprehensive identity securityCompany executed textbook SaaS transformation from perpetual licensing to subscription-based cloud modelLeadership set clear customer expectations, framing SaaS shift as faster innovation deliveryAddressed customer concerns about cost predictability, security compliance, and data residency requirementsTechnical team implemented lift-and-shift architecture with AWS RDS and multi-tenant improvementsCorporate initiative tracked weekly metrics and milestones throughout full development lifecycle processCustomer Success evolved from transactional support to strategic partnership embedded in security journeysAWS partnership fundamental to cloud journey with 25+ integrations and Marketplace collaborationAI strategy focuses on three pillars: using AI, securing against AI threatsFuture 12-24 months: continue securing all identities while expanding AI capabilities and solutionsAWS partnership expanding in 2025 leveraging machine identity leadership and GenAI advancesParticipants:Peretz Regev – Chief Product & Technology Officer, CyberArkBoaz Ziniman – Principal Developer Advocate - EMEA, Amazon Web ServicesFurther Links:· CyberArk: Website – LinkedIn – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

How do you decide whether to buy a ready-made AI solution or build one from the ground up? The choice is more than just a technical decision; it's about aligning AI with your business goals.   In this episode, Lois Houston and Nikita Abraham are joined by Principal Instructor Yunus Mohammed to examine the critical factors influencing the buy vs. build debate. They explore real-world examples where businesses must weigh speed, customization, and long-term strategy. From a startup using a SaaS chatbot to a bank developing a custom fraud detection model, Yunus provides practical insights on when to choose one approach over the other.   AI for You: https://mylearn.oracle.com/ou/course/ai-for-you/152601/   Oracle University Learning Community: https://education.oracle.com/ou-community   LinkedIn: https://www.linkedin.com/showcase/oracle-university/   X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.   ---------------------------------------------------------------   Episode Transcript: 00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:26 Nikita: Welcome to the Oracle University Podcast! I'm Nikita Abraham, Team Lead: Editorial Services with Oracle University, and with me is Lois Houston, Director of Innovation Programs. Lois: Hi there! Last week, we spoke about the key stages in a typical AI workflow and how data quality, feedback loops, and business goals influence AI success. 00:50 Nikita: In today's episode, we're going to explore whether you should buy or build AI apps. Joining us again is Principal Instructor Yunus Mohammed. Hi Yunus, let's jump right in. Why does the decision of buy versus build matter? Yunus: So when we talk about buy versus build matters, we need to consider the strategic business decisions over here. They are related to the strategic decisions which the business makes, and it is evaluated in the decision lens. So the center of the decision lens is the business objective, which identifies what are we trying to solve. Then evaluate our constraints based on that particular business objective like the cost, the time, and the talent. And finally, we can decide whether we need to buy or build. But remember, there is no single correct answer. What's right for one business may not be working for the other one. 01:54 Lois: OK, can you give us examples of both approaches? Yunus: The first example where we have got a startup using a SaaS AI chatbot. Now, being a startup, they have to choose a ready-made solution, which is an AI chatbot. Now, the question is, why did they do this? Because speed and simplicity mattered more than deep customization that is required for the chatbot. So, their main aim was to have it ready in short period of time and make it more simpler. And this actually lead them to get to the market fast with low upfront cost and minimal technical complexities. But in some situations, it might be different. Like, your bank, which needs to build a fraud model. It cannot be outsourced or got from the shelf. So, they build a custom model in-house. With this custom model, they actually have a tighter control, and it is tuned to their standards. And it is created by their experts. So these two generic examples, the chatbot and the fraud model example, helps you in identifying whether I should go for a SaaS product with simple choice of selecting an existing LLM endpoint and not making any changes. Or should I go with model depending on my business and organization requirement and fine tuning that model later to define a better implementation of the scenarios or conditions that I want to do which are specific to my organization. So here you decide with the reference whether I want it to be done faster, or whether I want to be more customized to my organization. So buy it, when it is generic, or build when it is strategic. The SaaS, which is basically software as a service, refers to ready to use cloud-based applications that you access via internet. You can log into the platform and use the built-in AI, there's no setup requirement for those. Real-world examples can be Oracle Fusion apps with AI features enabled. So in-house integration means embedding AI with my own requirements into your own systems, often using custom APIs, data pipelines, and hosting it. It gives you more flexibility but requires a lot of resources and expertise. So real-world example for this scenario can be a logistics heavy company, which is integrating a customer support model into their CX. 04:41 Lois: But what are the pros and cons of each approach? Yunus: So, SaaS and Fusion Applications, basically, they offer fast deployment with little to no coding required, making them ideal for business looking to get started quickly and faster. And they typically come with lower upfront costs and are maintained by vendor, which means updates, security, support are handled externally. However, there are limited customizations and are best suited for common, repeatable use cases. Like, it can be a standard chatbot, or it can be reporting tools, or off the shelf analytics that you want to use. But the in-house or custom integration, you have more control, but it takes longer to build and requires a higher initial investment. The in-house or custom integration approach allows full customization of the features and the workflows, enabling you to design and tailor the AI system to your specific needs. 05:47 Nikita: If you're weighing the choice between buying or building, what are the critical business considerations you'd need to take into account? Yunus: So let's take one of the key business consideration which is time to market. If your goal is to launch fast, maybe you're a startup trying to gain traction quickly, then a prebuilt plug and play AI solution, for example, a chatbot or any other standard analytical tool, might be your best bet. But if you have time and you are aiming for precision, a custom model could be worth the wait. Prebuilt SaaS tools usually have lower upfront costs and a subscription model. It works with putting subscriptions. Custom solutions, on the other hand, may require a bigger investment upfront. In development, you require high talent and infrastructures, but could offer cost savings in the long run. So, ask yourself a question here. Is this AI helping us stand out in the market? If the answer is yes, you may want to build something which is your proprietary. For example, an organization would use a generic recommendation engine. It's a part of their secret sauce. Some use cases require flexibility, like you want to tailor the rules to match your specific risk criteria. So, under that scenarios, you will go for customizing. So, you will go with off the shelf solutions may not give you deep enough requirements that you want to evaluate. So, you get those and you try to customize those. You can go for customization of your AI features. The other important key business consideration is the talent and expertise that your organization have. So, the question that you need to ask in the organization is, do you have an internal team who is well versed in developing AI solutions for you? Or do you have access to one of the teams which can help you build your own proprietary products? If not, you'll go with SaaS. If you do have, then building could unlock greater control over your AI features and AI models. The next core component is your security and data privacy. If you're handling sensitive information, like for example, the health care or finance data, you might not want to send your data to the third-party tools. So in-house models offer better control over data security and compliance. When we leverage a model, it could be a prebuilt or custom model. 08:50 Oracle University is proud to announce three brand new courses that will help your teams unlock the power of Redwood—the next generation design system. Redwood enhances the user experience, boosts efficiency, and ensures consistency across Oracle Fusion Cloud Applications. Whether you're a functional lead, configuration consultant, administrator, developer, or IT support analyst, these courses will introduce you to the Redwood philosophy and its business impact. They'll also teach you how to use Visual Builder Studio to personalize and extend your Fusion environment. Get started today by visiting mylearn.oracle.com.  09:31 Nikita: Welcome back! So, getting back to what you were saying before the break, what are pre-built and custom models? Yunus: A prebuilt model is an AI solution that has already been trained by someone else, typically a tech provider. It can be used to perform a specific task like recognizing images, translating text, or detecting sentiments. You can think of it like buying a preassembled appliance. You plug it in, configure a few settings, and it's ready to use. You don't need to know how the internal parts work. You benefit from the speed, ease, and reliability of this particular model, which is a prebuilt model. But you can't easily change how it works under the hood. Whereas, a custom model is an AI solution that your organization designs and trains and tunes specifically for their business problems using their own data. You can think of it like designing your own suit. It takes more time and effort to create. It is built to your exact measurements and needs. And you have full control over how it performs and evolves. 10:53 Lois: So, when would you choose a pre-built versus a custom model? Yunus: Depending on speed, simplicity, control, and customization, you can decide on using a prebuilt or to create a custom model. Prebuilt models are like plug and play solutions. Think of tools like Google Translate for languages. OpenAI APIs for summarizing sentiment analysis or chatbots, they are quick to deploy, require low technical effort, great for getting started fast, but they also have limits. Customization is very minimal, and you may not be able to fine tune it to your specific tone or business logic. These work well when the problem is common and nonstrategic, like, scanning documents or auto tagging images. The custom-build model, on the other hand, is a model that is built from the ground up. Using your own data and objectives, they take longer, and they require technical expertise. But they offer precise control, full alignment with your business needs. And these are ideal when you are dealing with sensitive data, competitive workflows, highly specific customer interactions. For example, a bank may build a custom model which can be used for fraud detection, which can be tuned to their exact transaction standards and the patterns of their transactions. 12:37 Nikita: What if someone wants the best of both worlds?  Yunus: We've also got a hybrid approach. In hybrid approach, we actually talk about the adaptation of AI with a strategy which is termed as hybrid strategy. Many companies today don't start by building AI from scratch. Instead, they begin with prebuilt models, like using an API, which can be already performing tasks like summarizing, translating, or answering questions using generic knowledge. This set will help you in getting up and running quickly with a small level results. As your business matures, you can start to layer in your custom data. Think internal policies, frequently asked questions, or customer interactions. And then you can fine tune the model to behave the way your business needs it to behave. Now, your AI starts producing business-ready output, smarter, more relevant, and aligned with your tone, brand, or compliance needs.  13:45 Lois: Ok…let's think of AI deployment in the hybrid approach as following a pyramid or ladder like structure. Can you take us through the different levels?  Yunus: So, on the top, quick start, minimal setup, great for business automation, which can be used as a pilot use case. So, if I'm taking off the shelf APIs or platforms, they can be giving me a faster, less set of requirements, and they are basically acting like a pilot use. Later, you can add your own data or logic so you can add your data. You can fine tune or change your business logic. And this is where fine tuning and prompt engineering helps tailor the AI to your workflows and your language. And then at the end, which is at the bottom, you build your own model. It is reserved for core capabilities or competitive advantages where total control and differentiation matters in building that particular model. You don't need to go all in from one day. So, start with what is available, like, use an off shelf, API, or platform, customize as you grow. Build only when it gives you a true edge. This is what we call the best of both worlds, build and buy. 15:05 Lois: Thank you so much, Yunus, for joining us again. To learn more about the topics covered today, visit mylearn.oracle.com and search for the AI for You course. Nikita: Join us next week for another episode of the Oracle University Podcast where we discuss the Oracle AI stack and Oracle AI services. Until then, this is Nikita Abraham… Lois: And Lois Houston, signing off! 15:29 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.

Three leading ISV executives from Coveo, DTEX Systems and Honeycomb, reveal how companies with proprietary datasets are gaining unbeatable competitive advantages in the AI era and share real-world strategies how you have similar outcomes.Topics Include:Panel introduces three ISV leaders discussing data platform transformation for AIDTEX focuses on insider threats, Coveo on enterprise search, Honeycomb on observabilityCompanies with proprietary datasets gain strongest competitive advantage in AI transformationData gravity concept: LLMs learning from unique datasets create defensible business positionsCoveo maintains unified enterprise index with real-time content and access rights syncHoneycomb enables subsecond queries for analyzing logs, traces, and metrics at scaleMulti-tenant architectures balance shared infrastructure benefits with single-tenant data separationCoveo deployed 140,000 times last year using mostly multi-tenant, some single-tenant componentsDTEX scaled from thousands to hundreds of thousands endpoints after architectural transformationCapital One partnership taught DTEX how to break monolithic architecture into servicesApache Iceberg and open table formats enable interoperability without data duplicationHoneycomb built custom format following similar patterns with hot/cold storage tiersBusiness data catalogs become critical for AI agents understanding dataset contextMCP servers allow AI systems to leverage structured cybersecurity datasets effectivelyDTEX used Cursor with their data to identify North Korean threat actorsReal-time AI data needs balanced with costs using right models for jobsCaching strategies and precise context reduce expensive LLM inference calls unnecessarilySearch remains essential for enterprise AI to prevent hallucination and access informationROI measurement focuses on cost reduction, analyst efficiency, and measurable business outcomesKey takeaway: invest in data structure early, context is king, AI is just softwareParticipants:Sebastien Paquet - Vice President of AI Strategy, CoveoRajan Koo - CTO, DTEX SystemsPatrick King - Head of Data, Honeycomb.ioKP Bhat - Sr Solutions Architecture Leader- Analytics & AI, Amazon Web ServicesFurther Links:Coveo: Website – LinkedIn – AWS MarketplaceDTEX Systems: Website – LinkedIn – AWS MarketplaceHoneycomb.io: Website – LinkedIn – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Okta's CTO Bhawna Singh discusses AI adoption, innovation and the four critical identity patterns needed to build the trust that accelerates AI implementation.Topics Include:AI innovation races ahead while adoption lags due to trust and security concernsResearch shows 82% plan AI deployment but 61% of customers demand trust firstAI coding tools dramatically reduce development time, accelerating software delivery cyclesAI interaction evolved from ChatGPT conversations to autonomous headless agents working independentlyFuture envisions millions of agents making decisions and communicating without human oversightComplex data relationships emerge as agents access multiple dynamic sources simultaneouslyTrust fundamentally starts with identity - the foundation for all AI securityFour critical identity patterns needed: authentication, API security, user confirmation, and authorizationAuthentication ensures legitimate agents while token vaults enable secure agent-to-agent communicationAsynchronous user approval prevents rogue decisions like the recent database deletion incidentIndustry standards like MCP protocol establish minimum security guardrails for interoperabilityTrust accelerates AI adoption through security, accountability, and collaborative standard-building effortsParticipants:Bhawna Singh – CTO, Customer Identity, OktaSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

A panel discussion with AI industry leaders revealing how enterprises are scaling AI today, with predictions on coming breakthroughs for AI and the impact on Fortune 500 companies and beyond.Topics Include:Three technical leaders discuss production challenges: security, interoperability, and scaling agentic systemsPanelists represent Enkrypt (security), Anyscale (infrastructure), and CrewAI (agent orchestration platforms)Industry moving from flashy demos to dependable agents with real business outcomesBreakthrough examples include 70-page IRS form processing and multimodal workflow automationMultimodal data integration becoming crucial - incorporating video, audio, screenshots into decisionsLess than 10% of future applications expected to be text-onlyCompanies shifting from experimenting with individual models to deploying agent networksNeed for governance frameworks as enterprises scale to hundreds of agentsGrowing software stack complexity requires specialized infrastructure between applications and GPUsSecurity teams need centralized visibility across fragmented agent deployments across enterprisesExisting industry regulations apply to AI services - no special AI laws neededInteroperability standards debate: MCP gaining adoption while A2A seems premature solutionMCP shows higher API reliability than OpenAI tool calling for implementationsMultimodal systems more vulnerable to attacks but value proposition too high ignoreFortune 500 company automated price operations approval process using 630 brands data87% of enterprise customers deploy agents in private VPCs or on-premises infrastructureSpecialized AI systems needed to oversee other agents at machine speed scalesCost optimization through model specialization rather than always using most powerful modelsFuture learning may happen through context/prompting rather than traditional weight fine-tuningPredictions include AI meeting moderators and agents working autonomously for hoursParticipants:Robert Nishihara - Co-founder, AnyscaleJoão Moura - CEO, CrewAISahil Agarwal - Co-Founder & CEO, Enkrypt AIJillian D'Arcy - Sr. ISV Sales Leader, Amazon Web ServicesFurther Links:Anyscale – Website | LinkedIn | AWS MarketplaceCrewAI - Website | LinkedIn | AWS MarketplaceEnkrypt AI - Website | LinkedIn | AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

In this episode of the Pipeliners Podcast, we revisit our conversation with Clint Bodungen of ThreatGEN. The discussion focuses on the application of gamification and generative AI in professional training, specifically for enhancing cybersecurity and incident response exercises. The episode also explores a PHMSA-sponsored R&D project that is adapting these advanced technologies for the unique operational needs of the pipeline industry, highlighting the development of AI-driven, multiplayer training environments.   Visit PipelinePodcastNetwork.com for a full episode transcript, as well as detailed show notes with relevant links and insider term definitions.

Join Lois Houston and Nikita Abraham as they chat with Yunus Mohammed, a Principal Instructor at Oracle University, about the key stages of AI model development. From gathering and preparing data to selecting, training, and deploying models, learn how each phase impacts AI's real-world effectiveness. The discussion also highlights why monitoring AI performance and addressing evolving challenges are critical for long-term success.   AI for You: https://mylearn.oracle.com/ou/course/ai-for-you/152601/252500   Oracle University Learning Community: https://education.oracle.com/ou-community   LinkedIn: https://www.linkedin.com/showcase/oracle-university/   X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode.   --------------------------------------------------------------   Episode Transcript: 00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:25 Lois: Welcome to the Oracle University Podcast! I'm Lois Houston, Director of Innovation Programs with Oracle University, and with me is Nikita Abraham, Team Lead: Editorial Services. Nikita: Hey everyone! In our last episode, we spoke about generative AI and gen AI agents. Today, we're going to look at the key stages in a typical AI workflow. We'll also discuss how data quality, feedback loops, and business goals influence AI success. With us today is Yunus Mohammed, a Principal Instructor at Oracle University.  01:00 Lois: Hi Yunus! We're excited to have you here! Can you walk us through the various steps in developing and deploying an AI model?  Yunus: The first point is the collect data. We gather relevant data, either historical or real time. Like customer transactions, support tickets, survey feedbacks, or sensor logs. A travel company, for example, can collect past booking data to predict future demand. So, data is the most crucial and the important component for building your AI models. But it's not just the data. You need to prepare the data. In the prepared data process, we clean, organize, and label the data. AI can't learn from messy spreadsheets. We try to make the data more understandable and organized, like removing duplicates, filling missing values in the data with some default values or formatting dates. All these comes under organization of the data and give a label to the data, so that the data becomes more supervised. After preparing the data, I go for selecting the model to train. So now, we pick what type of model fits your goals. It can be a traditional ML model or a deep learning network model, or it can be a generative model. The model is chosen based on the business problems and the data we have. So, we train the model using the prepared data, so it can learn the patterns of the data. Then after the model is trained, I need to evaluate the model. You check how well the model performs. Is it accurate? Is it fair? The metrics of the evaluation will vary based on the goal that you're trying to reach. If your model misclassifies emails as spam and it is doing it very much often, then it is not ready. So I need to train it further. So I need to train it to a level when it identifies the official mail as official mail and spam mail as spam mail accurately.  After evaluating and making sure your model is perfectly fitting, you go for the next step, which is called the deploy model. Once we are happy, we put it into the real world, like into a CRM, or a web application, or an API. So, I can configure that with an API, which is application programming interface, or I add it to a CRM, Customer Relationship Management, or I add it to a web application that I've got. Like for example, a chatbot becomes available on your company's website, and the chatbot might be using a generative AI model. Once I have deployed the model and it is working fine, I need to keep track of this model, how it is working, and need to monitor and improve whenever needed. So I go for a stage, which is called as monitor and improve. So AI isn't set in and forget it. So over time, there are lot of changes that is happening to the data. So we monitor performance and retrain when needed. An e-commerce recommendation model needs updates as there might be trends which are shifting.  So the end user finally sees the results after all the processes. A better product, or a smarter service, or a faster decision-making model, if we do this right. That is, if we process the flow perfectly, they may not even realize AI is behind it to give them the accurate results.  04:59 Nikita: Got it. So, everything in AI begins with data. But what are the different types of data used in AI development?  Yunus: We work with three main types of data: structured, unstructured, and semi-structured. Structured data is like a clean set of tables in Excel or databases, which consists of rows and columns with clear and consistent data information. Unstructured is messy data, like your email or customer calls that records videos or social media posts, so they all comes under unstructured data.  Semi-structured data is things like logs on XML files or JSON files. Not quite neat but not entirely messy either. So they are, they are termed semi-structured. So structured, unstructured, and then you've got the semi-structured. 05:58 Nikita: Ok… and how do the data needs vary for different AI approaches?  Yunus: Machine learning often needs labeled data. Like a bank might feed past transactions labeled as fraud or not fraud to train a fraud detection model. But machine learning also includes unsupervised learning, like clustering customer spending behavior. Here, no labels are needed. In deep learning, it needs a lot of data, usually unstructured, like thousands of loan documents, call recordings, or scan checks. These are fed into the models and the neural networks to detect and complex patterns. Data science focus on insights rather than the predictions. So a data scientist at the bank might use customer relationship management exports and customer demographies to analyze which age group prefers credit cards over the loans. Then we have got generative AI that thrives on diverse, unstructured internet scalable data. Like it is getting data from books, code, images, chat logs. So these models, like ChatGPT, are trained to generate responses or mimic the styles and synthesize content. So generative AI can power a banking virtual assistant trained on chat logs and frequently asked questions to answer customer queries 24/7. 07:35 Lois: What are the challenges when dealing with data?  Yunus: Data isn't just about having enough. We must also think about quality. Is it accurate and relevant? Volume. Do we have enough for the model to learn from? And is my data consisting of any kind of unfairly defined structures, like rejecting more loan applications from a certain zip code, which actually gives you a bias of data? And also the privacy. Are we handling personal data responsibly or not? Especially data which is critical or which is regulated, like the banking sector or health data of the patients. Before building anything smart, we must start smart.  08:23 Lois: So, we've established that collecting the right data is non-negotiable for success. Then comes preparing it, right?  Yunus: This is arguably the most important part of any AI or data science project. Clean data leads to reliable predictions. Imagine you have a column for age, and someone accidentally entered an age of like 999. That's likely a data entry error. Or maybe a few rows have missing ages. So we either fix, remove, or impute such issues. This step ensures our model isn't misled by incorrect values. Dates are often stored in different formats. For instance, a date, can be stored as the month and the day values, or it can be stored in some places as day first and month next. We want to bring everything into a consistent, usable format. This process is called as transformation. The machine learning models can get confused if one feature, like example the income ranges from 10,000 to 100,000, and another, like the number of kids, range from 0 to 5. So we normalize or scale values to bring them to a similar range, say 0 or 1. So we actually put it as yes or no options. So models don't understand words like small, medium, or large. We convert them into numbers using encoding. One simple way is assigning 1, 2, and 3 respectively. And then you have got removing stop words like the punctuations, et cetera, and break the sentence into smaller meaningful units called as tokens. This is actually used for generative AI tasks. In deep learning, especially for Gen AI, image or audio inputs must be of uniform size and format.  10:31 Lois: And does each AI system have a different way of preparing data?  Yunus: For machine learning ML, focus is on cleaning, encoding, and scaling. Deep learning needs resizing and normalization for text and images. Data science, about reshaping, aggregating, and getting it ready for insights. The generative AI needs special preparation like chunking, tokenizing large documents, or compressing images. 11:06 Oracle University's Race to Certification 2025 is your ticket to free training and certification in today's hottest tech. Whether you're starting with Artificial Intelligence, Oracle Cloud Infrastructure, Multicloud, or Oracle Data Platform, this challenge covers it all! Learn more about your chance to win prizes and see your name on the Leaderboard by visiting education.oracle.com/race-to-certification-2025. That's education.oracle.com/race-to-certification-2025. 11:50 Nikita: Welcome back! Yunus, how does a user choose the right model to solve their business problem?  Yunus: Just like a business uses different dashboards for marketing versus finance, in AI, we use different model types, depending on what we are trying to solve. Like classification is choosing a category. Real-world example can be whether the email is a spam or not. Use in fraud detection, medical diagnosis, et cetera. So what you do is you classify that particular data and then accurately access that classification of data. Regression, which is used for predicting a number, like, what will be the price of a house next month? Or it can be a useful in common forecasting sales demands or on the cost. Clustering, things without labels. So real-world examples can be segmenting customers based on behavior for targeted marketing. It helps discovering hidden patterns in large data sets.  Generation, that is creating new content. So AI writing product description or generating images can be a real-world example for this. And it can be used in a concept of generative AI models like ChatGPT or Dall-E, which operates on the generative AI principles. 13:16 Nikita: And how do you train a model? Yunus: We feed it with data in small chunks or batches and then compare its guesses to the correct values, adjusting its thinking like weights to improve next time, and the cycle repeats until the model gets good at making predictions. So if you're building a fraud detection system, ML may be enough. If you want to analyze medical images, you will need deep learning. If you're building a chatbot, go for a generative model like the LLM. And for all of these use cases, you need to select and train the applicable models as and when appropriate. 14:04 Lois: OK, now that the model's been trained, what else needs to happen before it can be deployed? Yunus: Evaluate the model, assess a model's accuracy, reliability, and real-world usefulness before it's put to work. That is, how often is the model right? Does it consistently perform well? Is it practical in the real world to use this model or not? Because if I have bad predictions, doesn't just look bad, it can lead to costly business mistakes. Think of recommending the wrong product to a customer or misidentifying a financial risk.  So what we do here is we start with splitting the data into two parts. So we train the data by training data. And this is like teaching the model. And then we have got the testing data. This is actually used for checking how well the model has learned. So once trained, the model makes predictions. We compare the predictions to the actual answers, just like checking your answer after a quiz. We try to go in for tailored evaluation based on AI types. Like machine learning, we care about accuracy in prediction. Deep learning is about fitting complex data like voice or images, where the model repeatedly sees examples and tunes itself to reduce errors. Data science, we look for patterns and insights, such as which features will matter. In generative AI, we judge by output quality. Is it coherent, useful, and is it natural?  The model improves with the accuracy and the number of epochs the training has been done on.  15:59 Nikita: So, after all that, we finally come to deploying the model… Yunus: Deploying a model means we are integrating it into our actual business system. So it can start making decisions, automating tasks, or supporting customer experiences in real time. Think of it like this. Training is teaching the model. Evaluating is testing it. And deployment is giving it a job.  The model needs a home either in the cloud or inside your company's own servers. Think of it like putting the AI in place where it can be reached by other tools. Exposed via API or embedded in an app, or you can say application, this is how the AI becomes usable.  Then, we have got the concept of receives live data and returns predictions. So receives live data and returns prediction is when the model listens to real-time inputs like a user typing, or user trying to search or click or making a transaction, and then instantly, your AI responds with a recommendation, decisions, or results. Deploying the model isn't the end of the story. It is just the beginning of the AI's real-world journey. Models may work well on day one, but things change. Customer behavior might shift. New products get introduced in the market. Economic conditions might evolve, like the era of COVID, where the demand shifted and the economical conditions actually changed. 17:48 Lois: Then it's about monitoring and improving the model to keep things reliable over time. Yunus: The monitor and improve loop is a continuous process that ensures an AI model remains accurate, fair, and effective after deployment. The live predictions, the model is running in real time, making decisions or recommendations. The monitor performance are those predictions still accurate and helpful. Is latency acceptable? This is where we track metrics, user feedbacks, and operational impact. Then, we go for detect issues, like accuracy is declining, are responses feeling biased, are customers dropping off due to long response times? And the next step will be to reframe or update the model. So we add fresh data, tweak the logic, or even use better architectures to deploy the uploaded model, and the new version replaces the old one and the cycle continues again. 18:58 Lois: And are there challenges during this step? Yunus: The common issues, which are related to monitor and improve consist of model drift, bias, and latency of failures. In model drift, the model becomes less accurate as the environment changes. Or bias, the model may favor or penalize certain groups unfairly. Latency or failures, if the model is too slow or fails unpredictably, it disrupts the user experience. Let's take the loan approvals. In loan approvals, if we notice an unusually high rejection rate due to model bias, we might retrain the model with more diverse or balanced data. For a chatbot, we watch for customer satisfaction, which might arise due to model failure and fine-tune the responses for the model. So in forecasting demand, if the predictions no longer match real trends, say post-pandemic, due to the model drift, we update the model with fresh data.  20:11 Nikita: Thanks for that, Yunus. Any final thoughts before we let you go? Yunus: No matter how advanced your model is, its effectiveness depends on the quality of the data you feed it. That means, the data needs to be clean, structured, and relevant. It should map itself to the problem you're solving. If the foundation is weak, the results will be also. So data preparation is not just a technical step, it is a business critical stage. Once deployed, AI systems must be monitored continuously, and you need to watch for drops in performance for any bias being generated or outdated logic, and improve the model with new data or refinements. That's what makes AI reliable, ethical, and sustainable in the long run. 21:09 Nikita: Yunus, thank you for this really insightful session. If you're interested in learning more about the topics we discussed today, go to mylearn.oracle.com and search for the AI for You course.  Lois: That's right. You'll find skill checks to help you assess your understanding of these concepts. In our next episode, we'll discuss the idea of buy versus build in the context of AI. Until then, this is Lois Houston… Nikita: And Nikita Abraham, signing off! 21:39 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.

AI executives from Archer, Demandbase and Highspot and AWS reveal how they're tackling AI's biggest challenges—from securing data, managing regulatory changes and keeping humans in the loop.Topics Include:Three AI leaders introduce their companies: Archer, Demandbase and Highspot's approaches to enterprise AIDemandbase's data strategy: Customer data stays isolated, shared data requires consent, public sources fuel trainingGeographic complexity: AI compliance varies dramatically between Germany, US, Canada, and California regulationsHighSpot tackles sales bias: Granular questions replace generic assessments for more accurate rep evaluationsSBI framework applied to AI: Specific behavioral observations create better, more actionable sales coachingAI transparency through citations: Timestamped evidence lets managers verify AI feedback and catch hallucinationsArcher handles 20-30K monthly regulations: AI helps enterprises manage overwhelming compliance requirements at scaleTwo compliance types explained: Operational (common across companies) versus business-specific regulatory requirementsEU AI Act adoption: US companies embracing European framework for responsible AI governanceHuman oversight becomes mandatory: Expert-in-the-loop reviews ensure AI decisions remain correctable and auditableThe bigger AI risk: Companies face greater danger from AI inaction than AI adoptionAgentic AI security challenges: Data layers must enforce permissions before AI access, not afterAI agents need identity management: Same access controls apply whether human clicks or AI actsHuman oversight in high stakes: Chief compliance officers demand transparency and correction capabilitiesFuture challenge identified: 80% of enterprise data behind firewalls remains invisible to AI modelsParticipants:Kayvan Alikhani - Global Head of Engineering- Emerging Solutions, Archer Integrated Risk ManagementUmberto Milletti - Chief R&D Officer, DemandbaseOliver Sharp - Co-Founder & Chief AI Officer, HighspotBrian Shadpour - General Manager, Security, Amazon Web ServicesFurther Links:Archer Integrated Risk Management: Website – LinkedIn – AWS MarketplaceDemandbase: Website – LinkedIn – AWS MarketplaceHighspot: Website – LinkedIn – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

In a fascinating discussion, Rob McGrorty, Product Leader of Agents at Amazon AGI Lab, reveals how rapidly AI agents are evolving with corporate adoption exploding as companies race to deploy production agents and the challenges and advantages they're experiencing.Topics Include:GenAI adoption outpaces all previous tech waves, growing faster than computers or internetEarly adopters tackle complex tasks while newcomers still use basic text manipulation featuresAI models double their single-call task capabilities every seven months, exponentially increasing powerAccelerating progress makes yesterday's magic mundane, unlocking mass creativity and customer demandAgents represent natural evolution: chatbots answered questions, now agents autonomously accomplish tasksAmazon's browser agent finds apartments, maps distances, ranks options using multiple transit modesCorporate adoption exploded: 33% piloting agents in 2024, 67% moving to production nowTwo main agent types today: API calling with tool use, browser automationCurrent applications mirror "RPA 2.0" - form filling, data extraction, website QA testingFuture brings multi-agent systems, self-directing loops, and agent-to-agent negotiation scenariosMajor challenges: data privacy, oversight protocols, error responsibility, and ecosystem sustainabilityTechnical hurdles include real-time accuracy measurement, latency issues, and quality assurance frameworksParticipants:Rob McGrorty – Product Leader, Agents at Amazon AGI LabSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Gagan Singh of Elastic discuses how agentic AI systems reduce analyst burnout by automatically triaging security alerts, resulting in measurable ROI for organizationsTopics Include:AI breaks security silos between teams, data, and tools in SOCsAttackers gain system access; SOC teams have only 40 minutes to detect/containAlert overload causes analyst burnout; thousands of low-value alerts overwhelm teams dailyAI inevitable for SOCs to process data, separate false positives from real threatsAgentic systems understand environment, reason through problems, take action without hand-holdingAttack discovery capability reduces hundreds of alerts to 3-4 prioritized threat discoveriesAI provides ROI metrics: processed alerts, filtered noise, hours saved for organizationsRAG (Retrieval Augmented Generation) prevents hallucination by adding enterprise context to LLMsAWS integration uses SageMaker, Bedrock, Anthropic models with Elasticsearch vector database capabilitiesEnd-to-end LLM observability tracks costs, tokens, invocations, errors, and performance bottlenecksJunior analysts detect nation-state attacks; teams shift from reactive to proactive securityFuture requires balancing costs, data richness, sovereignty, model choice, human-machine collaborationParticipants:Gagan Singh – Vice President Product Marketing, ElasticAdditional Links:Elastic – LinkedIn - Website – AWS Marketplace See how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Join hosts Lois Houston and Nikita Abraham, along with Principal AI/ML Instructor Himanshu Raj, as they discuss the transformative world of Generative AI. Together, they uncover the ways in which generative AI agents are changing the way we interact with technology, automating tasks and delivering new possibilities.   AI for You: https://mylearn.oracle.com/ou/course/ai-for-you/152601/252500   Oracle University Learning Community: https://education.oracle.com/ou-community   LinkedIn: https://www.linkedin.com/showcase/oracle-university/   X: https://x.com/Oracle_Edu   Special thanks to Arijit Ghosh, David Wright, Kris-Ann Nansen, Radhika Banka, and the OU Studio Team for helping us create this episode. ------------------------------------------------------- Episode Transcript: 00:00 Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we'll bring you foundational training on the most popular Oracle technologies. Let's get started! 00:25 Lois: Welcome to the Oracle University Podcast! I'm Lois Houston, Director of Innovation Programs with Oracle University, and with me is Nikita Abraham, Team Lead of Editorial Services.   Nikita: Hi everyone! Last week was Part 2 of our conversation on core AI concepts, where we went over the basics of data science. In Part 3 today, we'll look at generative AI and gen AI agents in detail. To help us with that, we have Himanshu Raj, Principal AI/ML Instructor. Hi Himanshu, what's the difference between traditional AI and generative AI?  01:01 Himanshu: So until now, when we talked about artificial intelligence, we usually meant models that could analyze information and make decisions based on it, like a judge who looks at evidence and gives a verdict. And that's what we call traditional AI that's focused on analysis, classification, and prediction.  But with generative AI, something remarkable happens. Generative AI does not just evaluate. It creates. It's more like a storyteller who uses knowledge from the past to imagine and build something brand new. For example, instead of just detecting if an email is spam, generative AI could write an entirely new email for you.  Another example, traditional AI might predict what a photo contains. Generative AI, on the other hand, creates a brand-new photo based on description. Generative AI refers to artificial intelligence models that can create entirely new content, such as text, images, music, code, or video that resembles human-made work.  Instead of simple analyzing or predicting, generative AI produces something original that resembles what a human might create.   02:16 Lois: How did traditional AI progress to the generative AI we know today?  Himanshu: First, we will look at small supervised learning. So in early days, AI models were trained on small labeled data sets. For example, we could train a model with a few thousand emails labeled spam or not spam. The model would learn simple decision boundaries. If email contains, "congratulations," it might be spam. This was efficient for a straightforward task, but it struggled with anything more complex.  Then, comes the large supervised learning. As the internet exploded, massive data sets became available, so millions of images, billions of text snippets, and models got better because they had much more data and stronger compute power and thanks to advances, like GPUs, and cloud computing, for example, training a model on millions of product reviews to predict customer sentiment, positive or negative, or to classify thousands of images in cars, dogs, planes, etc.  Models became more sophisticated, capturing deeper patterns rather than simple rules. And then, generative AI came into the picture, and we eventually reached a point where instead of just classifying or predicting, models could generate entirely new content.  Generative AI models like ChatGPT or GitHub Copilot are trained on enormous data sets, not to simply answer a yes or no, but to create outputs that look and feel like human made. Instead of judging the spam or sentiment, now the model can write an article, compose a song, or paint a picture, or generate new software code.  03:55 Nikita: Himanshu, what motivated this sort of progression?   Himanshu: Because of the three reasons. First one, data, we had way more of it thanks to the internet, smartphones, and social media. Second is compute. Graphics cards, GPUs, parallel computing, and cloud systems made it cheap and fast to train giant models.  And third, and most important is ambition. Humans always wanted machines not just to judge existing data, but to create new knowledge, art, and ideas.   04:25 Lois: So, what's happening behind the scenes? How is gen AI making these things happen?  Himanshu: Generative AI is about creating entirely new things across different domains. On one side, we have large language models or LLMs.  They are masters of generating text conversations, stories, emails, and even code. And on the other side, we have diffusion models. They are the creative artists of AI, turning text prompts into detailed images, paintings, or even videos.  And these two together are like two different specialists. The LLM acts like a brain that understands and talks, and the diffusion model acts like an artist that paints based on the instructions. And when we connect these spaces together, we create something called multimodal AI, systems that can take in text and produce images, audio, or other media, opening a whole new range of possibilities.  It can not only take the text, but also deal in different media options. So today when we say ChatGPT or Gemini, they can generate images, and it's not just one model doing everything. These are specialized systems working together behind the scenes.  05:38 Lois: You mentioned large language models and how they power text-based gen AI, so let's talk more about them. Himanshu, what is an LLM and how does it work?  Himanshu: So it's a probabilistic model of text, which means, it tries to predict what word is most likely to come next based on what came before.  This ability to predict one word at a time intelligently is what builds full sentences, paragraphs, and even stories.  06:06 Nikita: But what's large about this? Why's it called a large language model?   Himanshu: It simply means the model has lots and lots of parameters. And think of parameters as adjustable dials the model fine tuned during learning.  There is no strict rule, but today, large models can have billions or even trillions of these parameters. And the more the parameters, more complex patterns, the model can understand and can generate a language better, more like human.  06:37 Nikita: Ok… and image-based generative AI is powered by diffusion models, right? How do they work?  Himanshu: Diffusion models start with something that looks like pure random noise.  Imagine static on an old TV screen. No meaningful image at all. From there, the model carefully removes noise step by step to create something more meaningful and think of it like sculpting a statue. You start with a rough block of stone and slowly, carefully you chisel away to reveal a beautiful sculpture hidden inside.  And in each step of this process, the AI is making an educated guess based on everything it has learned from millions of real images. It's trying to predict.   07:24 Stay current by taking the 2025 Oracle Fusion Cloud Applications Delta Certifications. This is your chance to demonstrate your understanding of the latest features and prove your expertise by obtaining a globally recognized certification, all for free! Discover the certification paths, use the resources on MyLearn to prepare, and future-proof your skills. Get started now at mylearn.oracle.com.  07:53 Nikita: Welcome back! Himanshu, for most of us, our experience with generative AI is with text-based tools like ChatGPT. But I'm sure the uses go far beyond that, right? Can you walk us through some of them?  Himanshu: First one is text generation. So we can talk about chatbots, which are now capable of handling nuanced customer queries in banking travel and retail, saving companies hours of support time. Think of a bank chatbot helping a customer understand mortgage options or virtual HR Assistant in a large company, handling leave request. You can have embedding models which powers smart search systems.  Instead of searching by keywords, businesses can now search by meaning. For instance, a legal firm can search cases about contract violations in tech and get semantically relevant results, even if those exact words are not used in the documents.  The third one, for example, code generation, tools like GitHub Copilot help developers write boilerplate or even functional code, accelerating software development, especially in routine or repetitive tasks. Imagine writing a waveform with just a few prompts.  The second application, is image generation. So first obvious use is art. So designers and marketers can generate creative concepts instantly. Say, you need illustrations for a campaign on future cities. Generative AI can produce dozens of stylized visuals in minutes.  For design, interior designers or architects use it to visualize room layouts or design ideas even before a blueprint is finalized. And realistic images, retail companies generate images of people wearing their clothing items without needing real models or photoshoots, and this reduces the cost and increase the personalization.  Third application is multimodal systems, and these are combined systems that take one kind of input or a combination of different inputs and produce different kind of outputs, or can even combine various kinds, be it text image in both input and output.  Text to image It's being used in e-commerce, movie concept art, and educational content creation. For text to video, this is still in early days, but imagine creating a product explainer video just by typing out the script. Marketing teams love this for quick turnarounds. And the last one is text to audio.  Tools like ElevenLabs can convert text into realistic, human like voiceovers useful in training modules, audiobooks, and accessibility apps. So generative AI is no longer just a technical tool. It's becoming a creative copilot across departments, whether it's marketing, design, product support, and even operations.  10:42 Lois: That's great! So, we've established that generative AI is pretty powerful. But what kind of risks does it pose for businesses and society in general?  Himanshu: The first one is deepfakes. Generative AI can create fake but highly realistic media, video, audios or even faces that look and sound authentic.  Imagine a fake video of a political leader announcing a policy, they never approved. This could cause mass confusion or even impact elections. In case of business, deepfakes can be also used in scams where a CEO's voice is faked to approve fraudulent transactions.  Number two, bias, if AI is trained on biased historical data, it can reinforce stereotypes even when unintended. For example, a hiring AI system that favors male candidates over equally qualified women because of historical data was biased.  And this bias can expose companies to discrimination, lawsuits, brand damage and ethical concerns. Number three is hallucinations. So sometimes AI system confidently generate information that is completely wrong without realizing it.   Sometimes you ask a chatbot for a legal case summary, and it gives you a very convincing but entirely made up court ruling. In case of business impact, sectors like health care, finance, or law hallucinations can or could have serious or even dangerous consequences if not caught.  The fourth one is copyright and IP issues, generative AI creates new content, but often, based on material it was trained on. Who owns a new work? A real life example could be where an artist finds their unique style was copied by an AI that was trained on their paintings without permission.  In case of a business impact, companies using AI-generated content for marketing, branding or product designs must watch for legal gray areas around copyright and intellectual properties. So generative AI is not just a technology conversation, it's a responsibility conversation. Businesses must innovate and protect.  Creativity and caution must go together.   12:50 Nikita: Let's move on to generative AI agents. How is a generative AI agent different from just a chatbot or a basic AI tool?  Himanshu: So think of it like a smart assistant, not just answering your questions, but also taking actions on your behalf. So you don't just ask, what's the best flight to Vegas? Instead, you tell the agent, book me a flight to Vegas and a room at the Hilton. And it goes ahead, understands that, finds the options, connects to the booking tools, and gets it done.   So act on your behalf using goals, context, and tools, often with a degree of autonomy. Goals, are user defined outcomes. Example, I want to fly to Vegas and stay at Hilton. Context, this includes preferences history, constraints like economy class only or don't book for Mondays.  Tools could be APIs, databases, or services it can call, such as a travel API or a company calendar. And together, they let the agent reason, plan, and act.   14:02 Nikita: How does a gen AI agent work under the hood?  Himanshu: So usually, they go through four stages. First, one is understands and interprets your request like natural language understanding. Second, figure out what needs to be done, in this case flight booking plus hotel search.  Third, retrieves data or connects to tools APIs if needed, such as Skyscanner, Expedia, or a Calendar. And fourth is takes action. That means confirming the booking and giving you a response like your travel is booked. Keep in mind not all gen AI agents are fully independent.  14:38 Lois: Himanshu, we've seen people use the terms generative AI agents and agentic AI interchangeably. What's the difference between the two?  Himanshu: Agentic AI is a broad umbrella. It refers to any AI system that can perceive, reason, plan, and act toward a goal and may improve and adapt over time.   Most gen AI agents are reactive, not proactive. On the other hand, agentic AI can plan ahead, anticipate problems, and can even adjust strategies.  So gen AI agents are often semi-autonomous. They act in predefined ways or with human approval. Agentic systems can range from low to full autonomy. For example, auto-GPT runs loops without user prompts and autonomous car decides routes and reactions.  Most gen AI agents can only make multiple steps if explicitly designed that way, like a step-by-step logic flows in LangChain. And in case of agentic AI, it can plan across multiple steps with evolving decisions.  On the memory and goal persistence, gen AI agents are typically stateless. That means they forget their goal unless you remind them. In case of agentic AI, these systems remember, adapt, and refine based on goal progression. For example, a warehouse robot optimizing delivery based on changing layouts.  Some generative AI agents are agentic, like auto GPT. They use LLMs to reason, plan, and act, but not all. And likewise not all agentic AIs are generative. For example, an autonomous car, which may use computer vision control systems and planning, but no generative models.  So agentic AI is a design philosophy or system behavior, which could be goal-driven, autonomous, and decision making. They can overlap, but as I said, not all generative AI agents are agentic, and not all agentic AI systems are generative.  16:39 Lois: What makes a generative AI agent actually work?  Himanshu: A gen AI agent isn't just about answering the question. It's about breaking down a user's goal, figuring out how to achieve it, and then executing that plan intelligently. These agents are built from five core components and each playing a critical role.  The first one is goal. So what is this agent trying to achieve? Think of this as the mission or intent. For example, if I tell the agent, help me organized a team meeting for Friday. So the goal in that case would be schedule a meeting.  Number 2, memory. What does it remember? So this is the agent's context awareness. Storing previous chats, preferences, or ongoing tasks. For example, if last week I said I prefer meetings in the afternoon or I have already shared my team's availability, the agent can reuse that. And without the memory, the agent behaves stateless like a typical chatbot that forgets context after every prompt.  Third is tools. What can it access? Agents aren't just smart, they are also connected. They can be given access to tools like calendars, CRMs, web APIs, spreadsheets, and so on.  The fourth one is planner. So how does it break down the goal? And this is where the reasoning happens. The planner breaks big goals into a step-by-step plans, for example checking team availability, drafting meeting invite, and then sending the invite. And then probably, will confirm the booking. Agents don't just guess. They reason and organize actions into a logical path.  And the fifth and final one is executor, who gets it done. And this is where the action takes place. The executor performs what the planner lays out. For example, calling APIs, sending message, booking reservations, and if planner is the architect, executor is the builder.   18:36 Nikita: And where are generative AI agents being used?  Himanshu: Generative AI agents aren't just abstract ideas, they are being used across business functions to eliminate repetitive work, improve consistency, and enable faster decision making. For marketing, a generative AI agent can search websites and social platforms to summarize competitor activity. They can draft content for newsletters or campaign briefs in your brand tone, and they can auto-generate email variations based on audience segment or engagement history.  For finance, a generative AI agent can auto-generate financial summaries and dashboards by pulling from ERP spreadsheets and BI tools. They can also draft variance analysis and budget reports tailored for different departments. They can scan regulations or policy documents to flag potential compliance risks or changes.  For sales, a generative AI agent can auto-draft personalized sales pitches based on customer behavior or past conversations. They can also log CRM entries automatically once submitting summary is generated. They can also generate battlecards or next-step recommendations based on the deal stage.  For human resource, a generative AI agent can pre-screen resumes based on job requirements. They can send interview invites and coordinate calendars. A common theme here is that generative AI agents help you scale your teams without scaling the headcount.   20:02 Nikita: Himanshu, let's talk about the capabilities and benefits of generative AI agents.  Himanshu: So generative AI agents are transforming how entire departments function. For example, in customer service, 24/7 AI agents handle first level queries, freeing humans for complex cases.  They also enhance the decision making. Agents can quickly analyze reports, summarize lengthy documents, or spot trends across data sets. For example, a finance agent reviewing Excel data can highlight cash flow anomalies or forecast trends faster than a team of analysts.  In case of personalization, the agents can deliver unique, tailored experiences without manual effort. For example, in marketing, agents generate personalized product emails based on each user's past behavior. For operational efficiency, they can reduce repetitive, low-value tasks. For example, an HR agent can screen hundreds of resumes, shortlist candidates, and auto-schedule interviews, saving HR team hours each week.  21:06 Lois: Ok. And what are the risks of using generative AI agents?  Himanshu: The first one is job displacement. Let's be honest, automation raises concerns. Roles involving repetitive tasks such as data entry, content sorting are at risk. In case of ethics and accountability, when an AI agent makes a mistake, who is responsible? For example, if an AI makes a biased hiring decision or gives incorrect medical guidance, businesses must ensure accountability and fairness.  For data privacy, agents often access sensitive data, for example employee records or customer history. If mishandled, it could lead to compliance violations. In case of hallucinations, agents may generate confident but incorrect outputs called hallucinations. This can often mislead users, especially in critical domains like health care, finance, or legal.  So generative AI agents aren't just tools, they are a force multiplier. But they need to be deployed thoughtfully with a human lens and strong guardrails. And that's how we ensure the benefits outweigh the risks.  22:10 Lois: Thank you so much, Himanshu, for educating us. We've had such a great time with you! If you want to learn more about the topics discussed today, head over to mylearn.oracle.com and get started on the AI for You course.  Nikita: Join us next week as we chat about AI workflows and tools. Until then, this is Nikita Abraham…  Lois: And Lois Houston signing off!  22:32 That's all for this episode of the Oracle University Podcast. If you enjoyed listening, please click Subscribe to get all the latest episodes. We'd also love it if you would take a moment to rate and review us on your podcast app. See you again on the next episode of the Oracle University Podcast.  

Pete Rubio reveals how Rapid7 transformed to an AI-first platform that automates security investigations and accelerates results from hours to seconds.Topics Include:Pete Rubio introduces Rapid7's journey to becoming an AI-first cybersecurity platformCybersecurity teams overwhelmed by growing attack surfaces and constant alert fatigueCustomers needed faster response times, not just more alerts coming fasterLegacy tools created silos requiring manual triage that doesn't scale effectivelyAI must turn raw security data into real-time decisions humans can trustUnified data platform correlates infrastructure, applications, identity, and business context togetherAgentic AI automates investigative work, reducing analyst tasks from hours to secondsRapid7 evaluated multiple vendors, choosing AWS for performance, cost, and flexibilityNova models delivered unmatched performance for global scaling at controlled costsBedrock provided secure model deployment with governance and data privacy boundariesAWS partnership enabled co-development and rapid iteration beyond typical vendor relationshipsTransparent AI shows customers how models reach conclusions before automated actionsSOC analyst expertise continuously trains models with real-time security intelligenceGovernance frameworks and guardrails implemented from day one, not retrofitted laterFuture plans include customer AI integration and bring-your-own-model capabilitiesParticipants:Pete Rubio – Senior Vice President, Platform & Engineering, Rapid7Additional Links:Rapid 7 – LinkedIn - Website – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Panther CEO William Lowe explains how integrating Amazon Bedrock AI into their security platform delivered 50% faster alert resolution for enterprise customers while maintaining the trust and control that security practitioners demand.Topics Include:Panther CEO explains how Amazon partnership accelerates security outcomes for customersCloud-native security platform delivers 100% visibility across enterprise environments at scaleCustomers like Dropbox and Coinbase successfully replaced Splunk with Panther's solutionPlatform processes petabytes monthly with impressive 2.3-minute average threat detection timeCritical gap identified: alert resolution still takes 8 hours despite fast detectionSecurity teams overwhelmed by growing attack surfaces and severe talent burnoutConstant context switching across tools creates inefficiency and organizational collaboration problemsAI integration with Amazon Bedrock designed to accelerate security team decision-makingFour trust principles: verifiable actions, secure design, human control, customer data ownershipResults show 50% faster alert triage; future includes Slack integration and automationParticipants:· William H Lowe – CEO, PantherSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

AWS executives reveal how generative AI is fundamentally reshaping ISV business models, from pricing strategies to go-to-market approaches, and provide actionable insights for software companies navigating this transformation.Topics Include:Alayna Broaderson and Andy Perkins introduce AWS Infrastructure Partnerships and ISV SalesGenerative AI profoundly changing how ISVs build, deliver and market software productsTwo ISV categories emerging: established SaaS companies versus pure gen AI startupsLegacy SaaS firms struggle with infrastructure modernization and potential revenue cannibalizationPure gen AI companies face scaling challenges, reliability issues and cost optimizationRevenue models shifting from subscription-based to consumption-based pricing per token/prompt/taskFuture-proofing architecture critical as technology evolves rapidly like F-35 fighter jetsData becoming key differentiator, especially domain-specific datasets in healthcare and legalBalancing cost, accuracy, latency and customer experience creates complex optimization challengesMultiple specialized models replacing single solutions, with agentic AI accelerating this trendHuman capital challenges include retraining engineering teams and finding expensive AI talentSecurity, compliance and explainability now mandatory - no more black box solutionsEnterprise customers struggle with data organization and quantifying clear gen AI ROIISV pricing models evolving with tiered structures and targeted vertical use casesTraditional SaaS playbooks failing in generative AI landscape due to ROI uncertaintyPOC-based go-to-market with free trials and case study selling proving most effectivePricing strategies include AI gates, credit systems and separate SKUs for servicesCustomer trust requires proactive security messaging and auditable, transparent AI solutionsModular architecture enables evolution as new technologies emerge in fast-changing marketAWS positioning as ultimate gen AI toolkit partner with ISV collaboration opportunitiesParticipants:Alayna Broaderson - Sr Manager, Infrastructure Technology Partnership, Amazon Web ServicesAndy Perkins - General Manager, US ISV, Amazon Web ServicesSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Hear how PagerDuty and Zoom built successful AI products using Amazon Q-Index to solve real customer problems like incident response and meeting intelligence, while sharing practical lessons from their early adoption journey.Topics Include:David Gordon introduces AWS Q-Business partnerships with PagerDuty and ZoomMeet Everaldo Aguiar: PagerDuty's Applied AI leader with academia and enterprise backgroundPaul Magnaghi from Zoom brings AI platform scaling experience from SeattleQ-Business launched over a year ago as managed generative AI servicePlatform enables agentic experiences: content discovery, analysis, and process automationBuilt on AWS Bedrock with enterprise guardrails and data source integrationPartners wanted backend capabilities but preferred their own UI and modelsQ-Index provides vector database functionality for ISV partner integrationsEveraldo explains PagerDuty's evolution from traditional ML to generative AI solutionsHistorical challenges: alert fatigue, noise reduction using machine learning approachesNew gen AI opportunities: incident context, relevant data surfacing, automated postmortemsEngineering teams faced learning curve with agents and high-latency user experiencesPaul discusses Zoom's existing AI: virtual backgrounds and voice isolation technologyAI Companion strategy focused on simplicity during complex generative AI adoptionProblem identified: valuable meeting conversations disappear after Zoom calls endCustomer feedback revealed need for enterprise data integration beyond basic summariesGoal: combine unstructured conversations with structured enterprise data seamlesslyPagerDuty Advanced provides agentic AI for on-call engineers during incidentsQ-Index integration accesses internal documentation: Confluence pages, runbooks, proceduresDemo shows Slack integration pulling relevant incident response documentation automaticallyAccess control lists ensure users see only data they're authorized to accessZoom's AI companion panel enables real-time meeting questions and summariesExample use cases: decision tracking, incident analysis, action item identificationAdvice for starting: standardize practices and create internal development templatesSingle data access point reduces legal and security evaluation overheadCenter of excellence approach helps teams move quickly across product divisionsCut through generative AI buzzwords to focus on real user valueFederated AWS Bedrock architecture provides model choice and flexibility meeting customersCustomer trust alignment between Zoom conversations and AWS data handlingGetting started: PagerDuty Advance available now, Zoom AI free with paid add-onsParticipants:Everaldo Aguiar – Senior Engineering Manager, Applied AI, PagerDutyPaul Magnaghi – Head of AI & ISV Go To Market, ZoomDavid Gordon - Global Business Development, Amazon Q for Business. Amazon Web ServicesFurther Links:PagerDuty Website, LinkedIn & AWS MarketplaceZoom Website, LinkedIn & AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Teleport Co-Founder and CEO Ev Kontsevoy discusses the security vs productivity trade-off that plagues growing companies and how Teleport's trusted computing model protects against the exponential growth of cybersecurity threats.Topics Include:Teleport CEO explains how to make infrastructure "nearly unhackable" through trusted computingTraditional security vs productivity trade-off: high security kills team efficiencyCompanies buy every security solution but still get told they're at riskWhy "crown jewels" thinking fails: computers should protect everything at scaleModern infrastructure has too many access paths to enumerate and secureApple's PCC specification shows trusted computing working in real production environmentsAI revolutionizes both offensive and defensive cybersecurity capabilities for everyone80% of companies can't guarantee they've removed all ex-employee accessIdentity fragmentation across systems creates anonymous relationships and security gapsHuman error probability grows exponentially as companies scale in three dimensionsYour laptop already demonstrates trusted computing: seamless access without constant loginsApple ecosystem shows device trust at scale through secure enclavesAI agents need trusted identities just like humans and machinesAWS marketplace partnership accelerates deals and provides strategic account insightsHire someone who understands partnership dynamics before starting with AWSGenerative AI will make identity attacks cheaper and faster than everSecurity responsibility shifting from IT teams to platform engineering teamsTeleport's "steady state invariant": infrastructure locked down except during authorized workTemporary access granted through tickets, then automatically revoked after completionLegacy systems and IoT devices require extending trust models beyond cloud-nativeParticipants:Ev Kontsevoy – Co-Founder and CEO, TeleportFurther Links:Teleport WebsiteTeleport AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Raj Koo, CTO of DTEX Systems, discusses how their enterprise-grade generative AI platform detects and disarms insider threats and enables them to stay ahead of evolving risks.Topics Include:Raj Koo, CTO of DTEX Systems, joins from Adelaide to discuss insider threat detectionDTEX evolved from Adelaide startup to Bay Area headquarters, serving Fortune 500 companiesCompany specializes in understanding human behavior and intention behind insider threatsMarket shifting beyond cyber indicators to focus on behavioral analysis and detectionRecent case: US citizen sold identity to North Korean DPRK IT workersForeign entities used stolen credentials to infiltrate American companies undetectedDTEX's behavioral detection systems helped identify this sophisticated identity theft operationGenerative AI becomes double-edged sword - used by both threat actors and defendersBad actors use AI for fake resumes and deepfake interviewsDTEX uses traditional machine learning for risk modeling, GenAI for analyst interpretationGoal is empowering security analysts to work faster, not replacing human expertiseAWS GenAI Innovation Center helped develop guardrails and usage boundaries for enterpriseChallenge: enterprises must follow rules while hackers operate without ethical constraintsDTEX gains advantage through proprietary datasets unavailable to public AI modelsAWS Bedrock partnership enables private, co-located language models for data securityPrivate preview launched February 2024 with AWS Innovation Center acceleration supportSoftware leaders should prioritize privacy-by-design from day one of GenAI adoptionFuture threat: information sharing shifts from files to AI-powered data queriesMonitoring who asks what questions of AI systems becomes critical security concernDTEX contributes to OpenSearch development while building vector databases for analysisParticipants:Rajan Koo – Chief Technology Officer, DTEX SystemsFurther Links:DTEX Systems WebsiteDTEX Systems AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Half of the internet is fake, bot traffic according to reports. Then we talk about how most of the top Twitch streamers were BUYING viewers to bilk advertisers. The dead internet is here. Watch this podcast episode on YouTube and all major podcast hosts including Spotify. CLOWNFISH TV is an independent, opinionated news and commentary podcast that covers Entertainment and Tech from a consumer's point of view. We talk about Gaming, Comics, Anime, TV, Movies, Animation and more. Hosted by Kneon and Geeky Sparkles. D/REZZED News covers Pixels, Pop Culture, and the Paranormal! We're an independent, opinionated entertainment news blog covering Video Games, Tech, Comics, Movies, Anime, High Strangeness, and more. As part of Clownfish TV, we strive to be balanced, based, and apolitical. Get more news, views and reviews on Clownfish TV News - https://news.clownfishtv.com/ On YouTube - https://www.youtube.com/c/ClownfishTV On Spotify - https://open.spotify.com/show/4Tu83D1NcCmh7K1zHIedvg On Apple Podcasts - https://podcasts.apple.com/us/podcast/clownfish-tv-audio-edition/id1726838629

Enterprise AI leaders from C3 AI, Resolve AI, and Scale AI reveal how Fortune 100 companies are successfully scaling agentic AI from pilots to production and share secrets for successful AI transformation.Topics Include:Panel introduces three AI leaders from Resolve AI, C3 AI, and Scale AIResolve AI builds autonomous site reliability engineers for production incident responseC3 AI provides full-stack platform for developing enterprise agentic AI workflowsScale AI helps Fortune 100 companies adopt agents with private data integrationMoving from AI pilots to production requires custom solutions, not shrink-wrap softwareSuccess demands working directly with customers to understand their specific workflowsAll enterprise AI solutions need well-curated access to internal data and resourcesSoftware engineering has permanently shifted to agentic coding with no going backAI agents rapidly improving in reasoning, tool use, and contextual understandingIndustry moving from simple co-pilots to agents solving complex multi-step problemsSpiros coins new concept: evolving from "systems of record" to "systems of knowledge"Democratized development platforms let enterprises declare their own agent workflowsSemantic business layers enable agents to understand domain-specific enterprise operationsTrust and observability remain major barriers to enterprise agent adoptionOversight layers essential for agents making longer-horizon autonomous business decisionsPerformance tracking and calibration systems needed like MLOps for reasoning chainsCEO-level top-down support required for successful AI transformation initiativesTraditional per-seat SaaS pricing models completely broken for agentic AI solutionsIndustry shifting toward outcome-based and work-completion pricing models insteadReal examples shared: agent collaboration in production engineering and sales automationParticipants:Nikhil Krishnan – SVP & Chief Technology Officer, Data Science, C3 AISpiros Xanthos – Founder and CEO, Resolve AIVijay Karunamurthy – Head of Engineering, Product and Design / Field Chief Technology Officer, Scale AIAndy Perkins – GM, US ISV Sales – Data, Analytics, GenAI, Amazon Web ServicesFurther Links:C3 – Website – AWS MarketplaceResolve AI – Website – AWS MarketplaceScale AI – Website – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Industry leaders from Automation Anywhere and AWS discuss how modern customer data collection has evolved, and practical strategies for implementing enterprise automation at scale.Topics Include:Automation Anywhere and AWS experts discuss modern enterprise automation strategiesTraditional profiting strategies may not work with today's changing business modelsCustomer data collection methods have evolved across multiple platforms significantlyModern verification processes include automated validation systems and streamlined timelinesBackground check automation is increasingly handled by AI-powered models and systemsStanford's "Wonder Bread" research paper introduced revolutionary enterprise process observation technologyWonder Bread demonstrated AI systems watching and automatically learning hospital workflowsThe technology can author workflows by observing real enterprise processesEnterprise Process Management built around observed behaviors shows promising resultsVerification challenges exist since Wonder Bread research isn't widely publicized yetProcess observation technology could transform how enterprises handle workflow creationSalesforce Wizard Interface dominates many current automation implementations in enterprisesSalesforce Agent Codes offer alternative approaches to traditional automation methodsAWS platform selection involves careful consideration of enterprise integration needsDemo implementations showcase real-world timeline expectations and deployment maturity levelsCurrent automation solutions have reached significant scale across various industriesWorkflow automation differs fundamentally from true agentic intelligence systems capabilitiesAgentic AI demonstrates autonomous decision-making beyond simple rule-based automation processesUnderstanding this distinction helps organizations choose appropriate technology approaches effectivelySession concludes with clarity on modern automation landscape and implementation strategiesParticipants:Pratyush Garikapati – Director of Products, Automation AnywhereSreenath Gotur – Snr Generative AI Specialist, Amazon Web ServicesFurther Links:Automation Anywhere websiteAutomation Anywhere – AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Making Tech Work for SMBs with Steve Massaro | DailyCyber 271 In this episode of DailyCyber, I'm joined by Steve Massaro, IT consultant, MSP owner, and strategic technologist with over 20 years of experience helping businesses succeed through managed services and secure IT leadership.Steve shares insight from his years building MSPs in Hawaii and Kansas—from scaling tech teams to keeping SMB clients safe in today's evolving threat landscape.

On this episode, host Paul W. Grimm speaks with Professor Maura R. Grossman about the fundamentals of artificial intelligence and its growing influence on the legal system. They explore what AI is (and isn't), how machine learning and natural language processing work, and the differences between traditional automation and modern generative AI. In layman's terms, they discuss other key concepts, such as supervised and unsupervised learning, reinforcement training, and deepfakes, and other advances that have accelerated AI's development. Finally, they address a few potential risks of generative AI, including hallucinations, bias, and misuse in court, which sets the stage for a deeper conversation about legal implications on the next episode, "To Trust or Not to Trust: AI in Legal Practice." ABOUT THE HOSTJudge Paul W. Grimm (ret.) is the David F. Levi Professor of the Practice of Law and Director of the Bolch Judicial Institute at Duke Law School. From December 2012 until his retirement in December 2022, he served as a district judge of the United States District Court for the District of Maryland, with chambers in Greenbelt, Maryland. Click here to read his full bio.

AWS's Mark Relph draws fascinating parallels between today's AI revolution and the 1900s agricultural mechanization that delivered 2,000% productivity gains, while exploring how agentic AI will fundamentally reshape every aspect of software business models.Topics Include:Mark Relph directs AWS's data and AI partner go-to-market strategy teamHis role focuses on making ISV partners a force multiplier for customer successPreviously ran go-to-market for Amazon Bedrock, AWS's fastest growing service everCurrent AI adoption pace exceeds even the early cloud computing boom yearsHistorical parallel: 1900s agricultural mechanization delivered 2,000% productivity gains and 95% resource reductionFirst commercial self-propelled farming equipment revolutionized entire economies and never looked back500 machines formed the "Harvest Brigade" during WWII, harvesting from Texas to CanadaMark has spoken to 600+ AWS customers about GenAI over two yearsOrganizations range from AI pioneers to those still "fending off pirates" internallyGenAI has become a phenomenal assistant within organizations for content and automationAWS's AI stack has three layers: infrastructure, Bedrock, and applicationsBottom layer provides complete control over training, inference, and custom applicationsMiddle layer Bedrock serves as the "operating system" for generative AI applicationsTop layer offers ready-to-use AI through Q assistants and productivity toolsAI systems are rapidly becoming more complex with multiple model chainsMany current "agents" are just really, really long prompts (Mark's hot take)Task-specific models are emerging as one size won't fit all use casesEvolution moves from human-driven AI to agent-assisted to fully autonomous agentsAgent readiness requires APIs that allow software to interact autonomouslyTraditional UIs become unnecessary when agents interface directly with systemsCore competencies shift when AI handles the actual "doing" of tasksSales and marketing must adapt to agents delivering outcomes autonomouslyGo-to-market strategies need complete rethinking for an agentic worldThe agentic age is upon us and AWS partners should shape the futureParticipants:Mark Relph – Director – Data & AI Partner Go-To-Market, Amazon Web ServicesSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Building on last week's DigiMarCon 2025 recap, Leslie Richards and Isabelle Horan unpack the evolution from traditional Search Engine Optimization (SEO) to Generative Engine Optimization (GEO) and what that means for law firms today.

Industry leaders from Celonis and AWS explain why 2025 marks the inflection point for agentic AI and how early adopters are gaining significant competitive advantages in efficiency and innovation.Topics Include:AWS's Cristen Hughes and Celonis's Jeff Naughton discuss AI agent transformationAndy Jassy declares AI agents will fundamentally change how we workThree key trends make AI agents practical: smarter models, longer tasks, cheaper costsAI now beats humans on complex benchmarks for the first time everClaude 3.7 cracked graduate-level reasoning where humans previously dominated completelyAI evolved from brief interactions to managing sustained multi-step complex workflowsProcessing costs plummeted 99.7% making enterprise-grade AI economically viable at scaleWe're transitioning from 2023's adaptation era to 2025's human-AI collaboration eraBy 2028, AI will suggest actions to humans rather than vice versaAgents are autonomous software that plan, act, and reason independently with minimal interventionAgent workflow: receive human request, create plan, execute actions, review, adjust, deliverFour agent components: brain (LLM), memory (context), actions (tools), persona (role definition)AWS offers three building approaches: ready-made solutions, managed platform, DIY developmentKey enterprise applications: software development acceleration, customer care automation, knowledge work optimizationManual processes like accounts payable offer huge transformation opportunities through intelligent automationDeep process analysis is critical before deploying agents for maximum effectivenessCelonis pioneered process mining to help enterprises understand their actual workflow realitiesCompanies are collections of interacting processes that agents need proper context to navigateProcess intelligence provides agents with placement guidance, data feeds, monitoring, and workflow directionCelonis-AWS partnership demonstrates order management agents that automatically handle at-risk situationsParticipants:Jeff Naughton – SVP and Fellow, CelonisCristen Hughes – Solutions Architecture Leader, ISV, North America, Amazon Web ServicesFurther Links:Celonis WebsiteCelonis on AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Justin DiPietro, Co-Founder & Chief Strategy Officer of Glia, shares how they are leveraging AI to enhance the customer experience in the highly regulated world of financial institutions.Topics Include:Glia provides voice, digital, and AI services for customer-facing and internal operationsBuilt on "channel-less architecture" unlike traditional contact centers that added channels sequentiallyOne interaction can move seamlessly between channels (voice, chat, SMS, social)AI applies across all channels simultaneously rather than per individual channel700 customers, primarily banks and credit unions, 370 employees, headquartered in New YorkTargets 3,500 banks and credit unions across the United States marketFocuses exclusively on financial services and other regulated industriesAI for regulated industries requires different approach than non-regulated businessesTraditional contact centers had trade-off between cost and quality of serviceAI enables higher quality while simultaneously decreasing costs for contact centersNumber one reason people call banks: "What's my balance?" (20% of calls)Financial services require 100% accuracy, not 99.999% due to trust requirementsUses AWS exclusively for security, reliability, and future-oriented technology accessReal-time system requires triple-hot redundancy; seconds matter for live callsWorks with Bedrock team; customers certify Bedrock rather than individual featuresShowed examples of competitors' AI giving illegal million-dollar loans at 0%"Responsible AI" separates probabilistic understanding from deterministic responses to customersUses three model types: client models, network models, and protective modelsTraditional NLP had 50% accuracy; their LLM approach achieves 100% understandingPolicy is "use Nova unless" they can't, primarily for speed benefitsParticipants:Justin DiPietro – Co-Founder & Chief Strategy Officer, GliaFurther Links:Glia WebsiteGlia AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Ed Bailey, Field CISO at Cribl, shares how Cribl and AWS are helping customers rethink their data strategy by making it easier to modernize, reduce complexity, and unlock long-term flexibility.Topics Include:Ed Bailey introduces topic: bridging gap between security data requirements and budgetCompanies face mismatch: 10TB data needs vs 5TB licensing budget constraintsData volumes growing exponentially while budgets remain relatively flat year-over-yearIT security data differs from BI: enormous volume, variety, complexityMany companies discover 600+ data sources during SIEM migration projects50% of SIEM data remains un-accessed within 90 days of ingestionComplex data collection architectures break frequently and require excessive maintenanceTeams spend 80% time collecting data, only 20% analyzing for valueData collection and storage are costs; analytics and insights provide business valuePoor data quality creates operational chaos requiring dozens of browser tabsSOC analysts struggle with context switching across multiple disconnected systemsTraditional vendor approach: "give us all data, we'll solve problems" is outdatedData modernization requires sharing information widely across organizational business unitsData maturity model progression: patchwork → efficiency → optimization → innovationData tiering strategy: route expensive SIEM data vs cheaper data lake storageSIEM costs ~$1/GB while data lakes cost ~$0.15-0.20/GB for storageCompliance retention data should go to object storage at penny fractionsDecouple data retention from vendor tools to enable migration flexibilityCribl platform offers integrated solutions: Stream, Search, Lake, Edge componentsCustomer success: Siemens reduced 5TB to 500GB while maintaining security effectivenessParticipants:Edward Bailey – Field CISO, CriblFurther Links:Cribl WebsiteCribl on AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

The Company designs, develops, and operatesdigital infrastructure that transforms surplus renewable energy into globalcomputing resources. Soluna's pioneering data centers are strategicallyco-located with wind, solar, or hydroelectric power plants to supporthigh-performance computing applications, including Bitcoin Mining, GenerativeAI, and other compute-intensive applications. Soluna's proprietary softwareMaestroOS(™) helps energize a greener grid while delivering cost-effective andsustainable computing solutions and superior returns. A leading developer of green data centers that convert excess renewableenergy into global computing resources. Soluna builds modular, scalable datacenters for computing intensive, batchable applications such as Bitcoin mining,AI, and machine learning. Soluna provides a cost-effective alternative tobattery storage or transmission lines. Up to 30% of the power of renewableenergy projects can go to waste. Soluna's data centers enable clean electricityasset owners to ‘Sell. Every. Megawatt.'

Leslie Richards and Isabelle Horan share key takeaways from the 2025 DigiMarCon Conference, including how law firms can move beyond basic content generation and instead use generative AI tools for advanced strategic applications.

Sam Johnson, Chief Customer Officer of Jamf, discusses the implementation of AI built on Amazon Bedrock that is a gamechanger in helping Jamf's 76,000+ customers scale their device management operations.Topics Include:Sam Johnson introduces himself as Chief Customer Officer from Jamf companyJamf's 23-year mission: help organizations succeed with Apple device managementCompany manages 33+ million devices for 76,000+ customers worldwide from MinneapolisJamf has used AI since 2018 for security threat detectionReleased first customer-facing generative AI Assistant just last year in 2024Presentation covers why, how they built it, use cases, and future plansJamf serves horizontal market from small business to Fortune 500 companiesChallenge: balance powerful platform capabilities with ease of use and adoptionAI could help get best of both worlds - power and simplicityAI also increases security posture and scales user capabilities significantlyCustomers already using ChatGPT/Claude but wanted AI embedded in productBuilt into product to reduce "doorway effect" of switching digital environmentsCreated small cross-functional team to survey land and build initial trailRest of engineering organization came behind to build the production highwayTeam needed governance layer with input from security, legal, other departmentsEvaluated multiple providers but ultimately chose Amazon Bedrock for three reasonsAWS team support, large community, and integration with existing infrastructureUses Lambda, DynamoDB, CloudWatch to support the Bedrock AI implementationAI development required longer training/validation phase than typical product featuresReleased "AI Assistant" with three skills: Reference, Explain, and Search capabilitiesParticipants:Sam Johnson – Chief Customer Officer, JamfFurther Links:Jamf.comJamf on AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Mark Stevens, SVP, Channels and Alliances, discusses how SecurityScorecard's strategic partnership with AWS enables them to scale their security solutions through cloud infrastructure, marketplace integration, and co-sell programsTopics Include:SecurityScorecard founded 10 years ago to understand third-party vendor security postureCompany has grown to 3,000 enterprise customers and 200+ partners globallyEvolved from ratings to "supply chain detection and response" over last yearSupply chain threats have doubled, creating extended attack surfaces for companiesMany organizations don't know their vendor count or vulnerabilities within supply chainsSecurityScorecard provides visibility into attack surfaces and management tools for controlGenerative AI is central to their ecosystem, leveraging AWS Bedrock extensivelyThey scan the entire internet every two days at massive scaleHave scored 12 million companies with security scorecards to dateAll workloads run on AWS cloud infrastructure as their primary platformAWS partnership provides necessary scale for managing hundreds of thousands of vendorsCase study: Identified vendor misconfigurations that could shut down 1,000 locationsOwn massive 10-year data lake with tens of millions of companiesNew managed service combines AI automation with human analysts for supportLarge organizations cannot fully automate supply chain security management yetQuality threat intelligence data now valuable to SOC teams, not just riskThird-party risk management and SOC teams are slowly converging for better securityAWS marketplace integration provides frictionless customer experience and larger dealsCo-sell programs with AWS enterprise sales teams create effective flywheel motionFuture expansion includes identity management, response actions, and internal signal managementParticipants:Mark Stevens – SVP, Channels and Alliances, SecurityScorecardFurther Links:SecurityScorecard.ioSecurityScorecard AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Our U.S. Media & Entertainment Analyst Benjamin Swinburne discusses how GenAI is transforming content creation, distribution and also raising some serious ethical questions. Read more insights from Morgan Stanley.----- Transcript -----Welcome to Thoughts on the Market. I'm Ben Swinburne, Morgan Stanley's U.S. Media and Entertainment Analyst. Today – GenAI is poised to shake up the entertainment business. It's Wednesday, July 23, at 10am in New York.It's never been easier to create art for anyone – with a little help from GenerativeAI. You can transform photos of yourself or loved ones in the style of a popular Japanese movie studio or any era of visual art to your liking. You can create a short movie by simply typing in a few prompts. Even I can speak to youin several different languages. I can ask about the weather:Hvordan er været i dag?Wie ist das wetter heute?आज मौसम कैसा है? In the media and entertainment industry, GenAI is expected to bring about a seismic shift in how content is made and consumed. A recent production used AI to de-age actors and recreate the likeness of a deceased performer—cutting what used to take hundreds of VFX artists a year to just a few months with a small team. There are many other examples of how GenAI is revolutionizing how stories are told, from scriptwriting and editing to visual effects and dubbing. In music, GenAI is helping music labels identify emerging talent and generate new compositions. GenAI can even create songs using the voices of long-gone artists – potentially extending revenue far beyond an artist's lifetime. GenAI-driven tools have the potential to reduce TV and film production costs by 10–30 percent, with animation and post-production among the biggest savings opportunities. GenAI could also transform how content reaches audiences. Recommendation engines can become even more predictive, using behavioral data to serve up exactly what listeners want—sometimes before we know what we want. And there's more studios can achieve in post production. GenAI can already dub content in multiple languages, even syncing mouth movements to match the new dialogue. This makes global distribution faster, cheaper, and more culturally relevant. With better engagement comes better monetization. Platforms will use GenAI to introduce new pricing tiers, targeted advertising, and personalized superfan content that taps into niche audiences willing to pay more. But all this innovation brings up profound ethical concerns. First, there's the issue of consent and copyright. Can GenAI tools legally use an actor's name, likeness or voice? Then there's the question of authorship. If an AI writes a script or composes a song, who owns the rights? The creator or the GenAI model? Labor unions are understandably worried. In 2023, AI was a major sticking point in negotiations between Hollywood studios and writers' and actors' guilds. The fear? That AI could replace human jobs or devalue creative work. There are also legal battles. Multiple lawsuits are underway over whether AI models trained on copyrighted material without permission violate intellectual property laws. The outcomes of these cases could reshape the entire industry. But here's a big question no one can ignore: Will audiences care if content is AI-generated? Some consumers are fascinated by AI-created music or visuals, while others crave the emotional depth and authenticity that comes from human storytelling. Made-by-humans could become a premium label in itself. Now, despite GenAI's rapid rise, not every corner of entertainment is vulnerable. Live sports, concerts, and theater remain largely insulated from AI disruption. These experiences thrive on real-time emotion, unpredictability, and human connection—things AI can't replicate. In an AI-saturated world, the value of live events and sports rights will rise, favoring owners of sports rights and live platforms. So where do we go from here? By and large, we're entering an era where storytelling is no longer limited by budget or geography. GenAI is lowering the barriers to entry, expanding the creative class, and reshaping the economics of media. The winners in this new landscape will likely be companies that can scale—platforms with massive user bases, deep data pools, and the engineering talent to integrate GenAI seamlessly. But there's also room for agile newcomers who can innovate faster than the incumbents and disrupt the disrupters. No doubt, as the tools get better, the questions get harder. And that's where the real story begins. Thanks for listening. If you enjoy the show, please leave us a review wherever you listen and share Thoughts on the Market with a friend or colleague today.

Chief Product Development Officer Mitchell Johnson discusses how Sonatype protects enterprise developers from malicious open source components while keeping them productive through AI.Topics Include:Sonatype provides software supply chain solutions for enterprises using open source componentsThey serve large enterprises, government agencies, and critical infrastructure providers globallyMain challenge: keeping developers productive while maintaining secure software supply chainsCybercrime and supply chain attacks are massive, growing industries threatening developersAI adoption is happening faster than expected, profoundly changing development workflowsBad actors evolved from waiting for vulnerabilities to creating malicious componentsMalicious open source components specifically target developer and DevOps toolchainsSonatype's security research team uses AI/ML to analyze every open source componentThey can predict and block malicious components before entering customer environmentsAWS partnership helps Sonatype meet customers where they want to do businessPartnership focuses on go-to-market alignment, not just technical integrationAWS sales teams should be treated as extensions of your own sales organizationUnderstanding AWS sales structure and incentives is crucial for successful partnershipsAI development is following same pattern as open source adoption twenty years ago"Shadow AI" parallels the earlier "shadow IT" trend with open source softwareAI speeds up code generation but security review processes haven't kept paceDevelopers need a "Hippocratic Oath" - taking responsibility for AI-generated code outputWithin 24 months, professionals not skilled in AI will struggle to stay relevantSonatype's culture encourages curiosity, experimentation, and accepts failure as part of innovationTheir core mission: help developers focus on innovation, not security choresParticipants:Mitchell Johnson – Chief Product Development Officer, SonatypeFurther Links:Sonatype WebsiteSonatype on AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/

Founder and CTO Alex Rice discusses how HackerOne uses generative AI to automate security workflows and prioritizing accuracy over efficiency to achieve end-to-end outcomes.Topics Include:HackerOne uses ethical hackers and AI to find vulnerabilities before criminalsWhite hat hackers stress test systems to identify security weaknesses proactivelyGenerative AI plays a huge role in HackerOne's security operationsSecurity teams struggle with constant toil of finding and fixing vulnerabilitiesAI helps minimize toil through natural language interfaces and automationBoth good and bad actors have access to generative AI toolsSuccess requires measuring individual task inputs and outputs, not just aggregatesBreaking down workflows into granular tasks reveals measurable AI improvementsHackerOne deployed "Hive," their AI security agent to reduce customer toilInitial focus was on tasks where AI clearly outperformed humansStarted with low-hanging fruit before tackling more complex strategic workflowsAccuracy is the primary success metric, not just efficiency or speedSecurity requires precision; wrong fixes create bigger problems than inefficiencyCustomer acceptance and reduced time to remediation are north star metricsHumans remain the source of truth for validation and feedback loopsBreak down human jobs into granular AI tasks using systems thinkingBuild specific agents for individual tasks rather than entire job rolesKeep humans accountable for end-to-end outcomes to maintain customer trustAWS Bedrock chosen for security, confidentiality, and data separation requirementsMoving from efficiency improvements to entirely new AI-enabled capabilitiesParticipants:Alex Rice – Founder & CTO/CISO, HackerOneFurther Links:HackerOne WebsiteHackerOne on AWS MarketplaceSee how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon.com/isv/