The Tech Blog Writer Podcast

What happens when an organization writes careful AI governance policies but its infrastructure cannot enforce any of them? In this episode of Tech Talks Daily, I speak with Sabina Anja, Chief Technologist at Broadcom within the VMware Cloud Foundation division, about the infrastructure controls required as AI agents move from generating answers to accessing data, calling APIs, modifying systems, and triggering work. Sabina brings experience from both sides of enterprise technology. She remembers cabling networks, dealing with unstable infrastructure, and receiving those weekend calls when downtime had already upset the business. That background informs her belief that ambitious AI programs cannot succeed without stable, observable, and enforceable infrastructure beneath them. Many organizations are repeating a familiar pattern. Business teams adopt AI services before IT has established visibility, ownership, or control. The terminology may have changed from shadow IT to shadow AI, but the management problem remains. Sabina argues that CIOs first need an inventory of agents, nonhuman identities, data access, processes, and accountable owners. The risk becomes greater because agents behave differently from people. They operate across multiple systems at machine speed and can perform repeated actions without appreciating the wider business outcome. An agent does not need malicious intent to cause disruption. Excessive permissions, flat networks, inconsistent access rules, and years of deferred infrastructure work can give it plenty of opportunities. Sabina recommends brokered access rather than direct access, alongside dedicated virtual machines or namespaces, microsegmentation, lateral security, east-west policy controls, and tamper-evident logging. Organizations also need to define which data an agent can view, modify, or move, especially when sovereignty and regulatory requirements apply. One of Sabina's most memorable ideas is to treat an AI agent like a superhuman contractor. It should have a defined purpose, a named manager, a clear access specification, an activity record, and an end date. Additional permissions should be earned through evidence of reliable behavior rather than granted on the first day. She also warns about agent debt. AI systems are developing rapidly, so an agent created today may become outdated within months. Sabina recommends assuming that many agents will expire after six to nine months rather than allowing forgotten systems and permissions to accumulate indefinitely. For CIOs wanting an immediate test, her advice is straightforward. Create an inventory of nonhuman identities with production access. Then select one agent and examine every part of the infrastructure it attempted to reach. The question is not simply whether the application produced the expected result. Leaders should ask whether the agent entered systems, networks, or data stores that nobody expected it to access. We also challenge the familiar claim that AI agents will take everybody's jobs. Sabina sees an opportunity to remove repetitive tasks and give technology professionals new skills, although she warns that agents may behave like teenagers armed with infrastructure permissions. They may not take your job, but they could become remarkably good at testing your patience. I'd love to hear your thoughts. Does your organization know how many AI agents have production access and who is accountable for each one?  

What does a rising cloud bill actually tell you about the value your business is creating? Eight years after our first conversation, I welcome Kunal, co-founder and CEO of Unravel Data, back to Tech Talks Daily. We compare the data infrastructure he was optimizing during the Hadoop era with today's enterprise stacks built around Databricks, Snowflake, BigQuery, AI pipelines, and autonomous agents. Kunal says Unravel Data has analyzed over 10 billion workloads across hundreds of enterprises. From that work, he argues that data platforms and infrastructure can account for up to 60% of cloud spending at some global businesses, while 30% to 40% of data platform spending may produce no business value. These are company claims, but they frame a problem many technology and finance leaders will recognize. The cloud bill arrives after thousands of individual engineering decisions have already been made. We discuss where cloud waste hides, including oversized clusters, hot storage holding cold data, abandoned pipelines, inefficient queries, duplicate datasets, and development jobs consuming production-level resources. The people creating those workloads seldom see the price attached to their decisions, leaving technology leaders with an aggregated bill that explains what was purchased but not why it was needed. AI adds another complication. Humans create workloads at human speed, while agents can generate queries, launch infrastructure, and consume tokens around the clock. An agent is designed to complete its task, not worry about whether a single query costs $5 or $5,000. Kunal argues that machine-speed consumption cannot be governed through monthly human reviews. We also discuss the difference between cost cutting and cost optimization, why aggressive reductions can damage performance and reliability, and how FinOps must connect cost with business outcomes. Kunal explains why leaders should measure cost per pipeline, model, agent, successful run, customer report, and business result. Finally, we consider the benefits and risks of autonomous data platform optimization. Kunal describes autonomy as a dial, with bounded, reversible, and validated actions earning wider authority as trust develops. Does your cloud bill show healthy growth, or is expensive waste hiding behind the headline number? Share your thoughts with me.

How can an independent safety evaluation remain meaningful when the AI inside a product may change after its next update? In this episode of Tech Talks Daily, I speak with Dr. Robert Slone, Senior Vice President, Chief Scientist, and Innovation Officer at UL Solutions. Robert has spent almost 30 years leading science, research, product development, and innovation teams. He now helps guide UL Solutions' scientific work across safety, security, and sustainability. Many listeners will recognize the UL Mark without knowing what happens behind it. Robert explains how UL Solutions tests products to their limits, which can involve setting them on fire, finding their breaking points, inspecting manufacturing facilities, and determining whether they meet defined safety requirements. That work began over 130 years ago when electricity was introducing unfamiliar risks. Today, the same broad question applies to artificial intelligence: how can society benefit from a new technology while understanding and managing the harm it could cause? The need is becoming increasingly visible as AI moves into healthcare, transportation, manufacturing, financial services, infrastructure, and consumer products. Robert recalls being approached about evaluating an AI-enabled teddy bear capable of talking with children. It is a memorable example of how decisions made inside an AI model can reach directly into everyday life. Robert organizes AI product safety around three pillars. The technical pillar considers robustness, risk management, functional safety, and whether the system performs its intended purpose. The ethical pillar includes fairness, bias, privacy, transparency, and explainability. Governance covers data management, product updates, accountability, and the complete operating life of the system. We also discuss one of the hardest problems in AI certification. Traditional products and software can be evaluated against a defined version, but AI systems may be updated, retrained, or affected by changing data. Robert explains why meaningful safety assurance requires version-specific testing, annual reviews, disclosure of significant changes, and eventually telemetry capable of identifying problems much closer to real time. For business leaders buying AI, the conversation provides a practical vendor checklist. Where did the training data come from? How was performance measured? What are the system's known limitations? How were privacy and bias assessed? Who takes responsibility if its behavior changes? Independent testing cannot promise that an evolving product will remain safe forever. It can provide evidence about the version evaluated, expose gaps, establish accountability, and create a process for monitoring future changes. What proof would you demand before allowing an AI product to influence an employee, patient, customer, or child? Listen to the episode and share your thoughts with me.    

Could allowing patients to choose their own appointment times help reduce missed visits and shorten NHS waiting lists? In this episode of Tech Talks Daily, I speak with Alison MacDonald, European Lead and Senior Vice President at Nordic Global. Alison brings an unusual combination of clinical and technology experience as a registered nurse who moved into digital health over 15 years ago. Her career began in community nursing, where she was asked to lead an electronic health record project because colleagues thought she was good with computers. What initially appeared to be a simple exercise in converting paper forms into digital records encouraged her to question whether healthcare could redesign the process rather than copy it onto a screen. We discuss Nordic's work with Cambridge University Hospitals NHS Foundation Trust on patient self-scheduling and automated earlier appointment offers. Before the program, missed outpatient appointments were removing valuable clinical capacity while administrative teams spent time calling patients and rearranging bookings. Cambridge introduced self-scheduling through Epic MyChart, allowing patients to select appointment times through the patient portal. Alison says the DNA rate fell from 5% to 2.2% during the program. According to the supplied results, over 20,000 patients successfully scheduled their own appointments and over 3,000 accepted offers to attend earlier when cancellations created availability. Patients moved appointments forward by an average of 16 days. Forty percent of accepted earlier appointments occurred within seven days of the offer, while 7% took place on the same or following day. Administrative teams also saved an estimated ten minutes for every self-booked appointment. Alison explains why patient control can improve attendance. People can choose times that work around employment, caring responsibilities, travel, and family life instead of receiving a fixed appointment through a letter or telephone call. Patients can also cancel or reschedule without waiting for somebody to answer the phone. The operational lesson goes beyond appointment booking. Healthcare systems may be able to recover existing capacity by examining missed appointments, theater scheduling, waiting list processes, pre-visit questionnaires, and patient communications before concluding that every problem requires additional staff or facilities. We also discuss where AI is producing practical results in healthcare. Alison points to medical imaging, emergency department triage, waiting list management, clinical documentation, and workforce deployment. She warns against discussing AI as one generic solution because each application requires a defined use case, suitable data, workable processes, governance, and staff adoption. Ambient clinical documentation offers one example. An AI scribe can record a consultation, prepare a structured note, and pass it to the clinician for review and correction. Alison cites an NHS evaluation reporting a 23.5% increase in direct patient interaction time and an 8.2% reduction in appointment length. Interoperability remains another major challenge. Healthcare journeys cross hospitals, primary care, community services, and specialist providers that may use different systems or a mixture of electronic and paper records. Even basic differences, such as one organization measuring pain on a five-point scale and another using ten points, can prevent reliable comparison. Alison recommends agreeing on common data standards, defining the minimum patient information required during care transitions, including interoperability requirements in procurement, and avoiding bespoke integrations that make future information sharing harder. Digital access also requires balance. Online services can improve convenience, but healthcare providers must retain appropriate alternatives for patients who lack digital skills, connectivity, confidence, or access. Could your healthcare organization improve patient access and staff capacity by redesigning one familiar process before purchasing another large technology platform? Listen to the episode and share your thoughts with me.

What would change if a customer could return three days later through a different channel and continue the same conversation without repeating a single detail? In this episode of Tech Talks Daily, I speak with Paul Adams, Chief Product Officer at Fin, the company previously known as Intercom. Paul has spent almost 13 years with the business and provides a candid account of how it abandoned its previous roadmap, placed a company-wide bet on AI, and rebuilt its products and working practices around AI agents. Our conversation begins with Fin's move from a customer service agent toward what Paul calls a single customer agent. The idea is that customers do not care whether their request belongs to sales, service, or customer success. They want the company to understand their situation and help them complete the task. Paul explains how AI agents can bring customer history, company knowledge, operational data, and business goals into the same conversation. This could allow an agent to resolve an issue, support a purchase, recognize a valuable customer, or transfer the conversation to a human without losing the context already provided. We also examine the economics behind poor customer service. Many companies are not ignoring customers through a lack of concern. They are receiving volumes of requests that cannot economically be handled by adding people alone. Paul says some Fin customers are resolving between 70 and 90 percent of customer queries through AI. Rather than seeing entire teams disappear, he is observing employees move into customer success, knowledge management, AI supervision, and higher-touch services. The episode also provides an unusually candid account of what it took to rebuild an established SaaS company around AI. Paul describes the process as brutal. Strategies were discarded, familiar processes were removed, and some people decided the new direction was not for them. His advice is to prioritize speed, place smaller experiments in front of real customers, and learn from evidence rather than waiting for every internal condition to become perfect. Paul also recalls working on early versions of mobile YouTube and Gmail when colleagues questioned whether anyone would watch video or answer email on a phone. Those stories provide a timely warning about judging new technology by its early limitations. Could AI agents finally give customers continuity across sales, service, and support, or will internal company structures remain the greater obstacle? Listen to the conversation and share your thoughts with me.

Can organizations trust an AI recommendation when they cannot understand the evidence, relationships, and previous decisions behind it? In this episode of Tech Talks Daily, I welcome back Jim Webber, Chief Scientist at Neo4j, to discuss the company's acquisition of GraphAware and its move from graph database provider to graph intelligence platform. GraphAware has worked with Neo4j for many years and developed Hume, an intelligence analysis platform used to connect and examine complex information. Bringing the two companies together gives Neo4j a direct role in applications serving police forces, governments, intelligence agencies, and other organizations handling connected data. Jim explains why context has become one of the biggest requirements for dependable AI. Enterprises already possess enormous volumes of data, but facts alone provide endpoints rather than the complete path leading to a decision. An agent needs to understand the knowledge available, the conversation taking place, and the record of previous decisions. It also needs to know which actions produced good outcomes and which produced poor ones. Jim compares these information layers to SimCity. Each can be viewed separately, but their greater value appears when they are combined. Knowledge, conversations, and decision traces can then help an agent understand why something happened and learn from the result. This introduces an interesting lesson from scientific research. Positive outcomes are frequently published, while failed experiments receive less attention. An AI agent needs both. Recording the breadcrumbs behind good and bad decisions provides the material required to improve its future behavior. We also discuss why large language models cannot understand every organization by themselves. Jim describes a model as a lossy compression of the internet. It can generate impressive natural language, but it does not automatically understand a company's policies, customers, history, evidence, or operating environment. Retrieval augmented generation can introduce relevant organizational information into the process. Graph RAG adds relationships between facts, helping the system understand how people, events, products, accounts, and other entities connect. According to research Jim references from the National Innovation Centre for Data, Graph RAG can improve accuracy while reducing costs by using fewer, higher-quality tokens. Explainability becomes especially important when AI supports decisions across policing, cyber defense, taxation, intelligence, banking, and government. A fluent answer may sound authoritative while containing a serious technical mistake. Jim shares an example from his own work where an agent confidently warned him about a "committed minority" inside a fault-tolerant computing protocol. The statement sounded plausible, but only a majority could commit within that protocol. Someone without Jim's technical knowledge might have accepted the recommendation and removed working code. This leads us to human oversight. Jim argues that the correct level depends on the consequences of the action. Automating a routine banking process with monitoring and safeguards may improve the customer experience. Ordering someone's arrest based solely on an agent's conclusion demands human involvement. We also consider digital sovereignty and why control over data has become a strategic concern for governments and large enterprises. Geopolitical instability, overseas technology dependencies, privacy requirements, and changing national policies are forcing leaders to ask where their data resides and whether they can retrieve or move it. Jim explains how Neo4j intends to offer organizations flexibility over where their information is stored and how it is deployed. The discussion also examines the opportunity for Neo4j and Hume to provide an alternative within a market where Palantir has held a powerful position. Looking ahead, Jim imagines intelligence analysts directing swarms of digital agents. Those agents could search data, connect evidence, identify relevant patterns, and present findings while humans retain responsibility for consequential decisions. If AI can connect information at machine speed, how do we ensure the person making the final decision can inspect the evidence and challenge the conclusion? Listen to the episode and share your thoughts with me.

What does an AI agent need to understand about your business before you allow it to make decisions and take action without waiting for human approval? In this episode of Tech Talks Daily, I speak with Kash Mehdi, Field CTO at Reltio, about the move from analytical AI that supports decisions to agentic AI that can execute them. Kash argues that leaders should begin treating AI agents as a workforce rather than another collection of software tools. A digital workforce needs training, boundaries, oversight, trusted information, and clear permissions before it can act safely. He uses the analogy of raising a puppy. When the puppy misbehaves, the problem may be inadequate training or poorly defined boundaries. AI agents present a similar leadership challenge. Organizations must ask what the agent has learned about the business and what authority it has been given. We discuss why model selection may be receiving too much executive attention. Kash describes four components of an agentic system: the model, tools, data, and context. Models are improving rapidly and tools are increasingly available, but business context remains incomplete across many enterprises. Data tells an agent a fact. Context helps it understand what the fact means within a particular customer relationship, geography, policy, or business process. Kash illustrates the difference with a pizza order. The data may confirm that someone is logged in, the model can interpret the request, and a tool can place the order. Context tells the system that it is Friday night, the customer is watching television, and they usually order pineapple and cheese pizza. The same principle becomes far more serious when an agent is dealing with medical equipment, supply chains, financial customers, or regulated information. It must understand which entities exist, how they relate, what information it may access, and which actions it has authority to complete. Kash identifies three requirements for safer autonomy: a governed source of truth, a live feedback loop, and enforceable permission boundaries. Trust must be built into the data and operating rules before the agent acts because the familiar human review step may no longer exist. We also discuss how governance changes when AI can execute decisions at machine speed. A poor decision made by one employee can usually be reviewed and corrected. A poor decision repeated automatically across thousands or millions of transactions can become a business incident before anyone intervenes. Kash shares examples involving restaurant menu launches, medical equipment deliveries, and call center offers. Each depends on current information and the relationships connecting customers, products, suppliers, locations, and previous interactions. For CIOs preparing today, Kash recommends building context around reusable entities rather than constructing an isolated data project for every AI use case. He points to Schneider Electric as an example where one unified foundation supported sales, shipping, operations, and marketing use cases. The conversation ends with a warning about slow data. Autonomous agents need current context because information that arrives after a decision has been made may no longer carry much business value. Kash predicts that the half-life of enterprise data will become a board-level measure. If a smarter agent can make a poor decision faster and with greater confidence, is your organization investing enough in the context, governance, and feedback needed to keep it on course? Listen to the conversation and share your thoughts with me.     Useful Links https://www.reltio.com/ https://www.reltio.com/datadriven/  

How much of your technology stack is being held together by people swiveling between screens, copying information, and quietly compensating for systems that cannot communicate? In this episode, I speak with John Foreman, Chief Product Officer at Clio, about what he calls the "swivel chair problem." John previously served as Chief Product Officer at Mailchimp and Podium, and now helps guide product development at a company seeking to support the complete operation of a law firm. We discuss why legal professionals have moved from understandable caution around AI toward increasingly sophisticated daily use. John explains why concerns about client confidentiality, intellectual property, model training, and data access initially slowed adoption, as well as why lawyers are now helping set the pace for responsible professional AI use. Our conversation also examines why disconnected technology stacks make AI appear far less capable. People can interpret information across documents, billing platforms, case management tools, email, and court systems. An AI system cannot perform the same work unless it receives the necessary context and access. John also explains why the familiar chatbot may be the wrong interface for many jobs. Some AI tasks should happen quietly, while work involving legal filings and client records requires structured review, accountability, and human approval. With lawyers spending an average of 62% of their time on nonbillable work, the immediate opportunity could include intake, billing, timekeeping, reviews, document processing, and filing. These lessons extend well beyond legal services. Where is the swivel chair problem hiding inside your organization? Listen to the conversation and share your thoughts with me.

Could Making Tax Digital improve the way small businesses manage their finances, or will it become another administrative burden competing for an already crowded evening? In this episode, I speak with Caroline Duong, Head of Business Admin at ANNA Money, about Making Tax Digital, quarterly reporting, AI bookkeeping, and the reality of running a small business when one person is often responsible for almost everything. ANNA Money stands for Absolutely No Nonsense Admin. It is an AI-powered, app-based business account and financial admin service designed for small businesses, startups, freelancers, and sole traders in the UK. Its goal is to reduce the paperwork that regularly follows business owners home after the working day has supposedly ended. Caroline explains that Making Tax Digital quarterly updates are reports to HMRC rather than full tax returns. The intention is to encourage people with self-employment or property income to maintain digital records throughout the year instead of rebuilding their finances from receipts shortly before a deadline. Awareness remains a problem. Caroline says an estimated 864,000 people are expected to submit updates during the first year, while fewer than half had signed up at the time of recording. HMRC's softer first-year approach gives people time to adjust, but Caroline warns against waiting until penalties enter the system before changing established habits. We also discuss what AI can do differently from traditional accounting software. Caroline offers a wonderfully simple example: a tire purchase may represent vehicle maintenance for one business and inventory for a car parts dealer. An AI system with enough business context can recognize that difference and categorize the transaction accordingly. Caroline also explains why responsible automation still needs human confirmation. Software can learn about suppliers, customers, and regular expenses, but it must recognize when information is missing or a decision requires human judgment. The conversation ends with two practical recommendations. Keep business and personal transactions separate, and begin tracking income and expenses early. Both can make quarterly reporting significantly easier and reduce the risk of being caught off guard later. If AI can give business owners a few hours back each month, which administrative task should it take on first? Listen to the conversation and share your thoughts with me.

Who really controls your enterprise technology strategy: your organization or the vendors writing its software contracts? In this episode of Tech Talks Daily, I speak with Tomás O'Leary, founder and CEO of Origina, about enterprise software vendor lock in, forced upgrades, subscription contracts, and the financial consequences of surrendering control over mission-critical systems. Tomás founded Origina in Dublin after working within the enterprise software supply chain and questioning the value customers received from traditional support contracts. He saw organizations paying substantial annual fees while experiencing poor response times, constant pressure to change versions, and upgrades that produced limited business value. He argues that the balance of power between technology buyers and suppliers has moved heavily toward the vendor. Companies that previously purchased perpetual software rights are increasingly being encouraged or forced toward subscription models, while complex contract terms and audit risks can make customers feel trapped. Some Origina customers have described this behavior as a "digital mafia," while one Fortune 50 organization, according to Tomás, uses AI to assess whether suppliers could be acquired by vendors it considers predatory. That business then considers longer contracts as protection against future licensing changes. However, leaving a vendor does not always require replacing the software. Tomás explains why perpetual software rights and independent support can give companies another option. A system that continues to perform its required business function may not need to be replaced simply because the original vendor has ended support or introduced a new commercial model. We discuss how leaders should distinguish between technology that genuinely requires modernization and dependable systems of record that could continue operating securely. Payroll platforms, general ledgers, claims systems, and other back-office applications may not require constant reinvention if the business requirement remains stable. Tomás also describes a European organization spending approximately €1 million annually on a software product. The company estimated that a vendor-required version change would cost €30 million. By moving to an alternative support arrangement, it expects to defer that expenditure while keeping the existing system operational. These figures are the organization's estimates, shared by Tomás during our conversation. We also discuss centralized technology dependency, outages, software patching, AI-assisted development, and why some companies are returning to internally developed applications for operations they consider particularly important. Tomás recommends that CIOs create a small team combining technical, procurement, contractual, and legal knowledge. This group should remain close to senior leadership and challenge assumptions before renewals, migrations, or major software changes are approved. Is your organization modernizing because the business needs to change, or because a vendor has decided that time is up? Listen to the conversation and share your thoughts with me.

Why are companies preparing for agent-to-agent customer service when many customers still cannot get a chatbot to answer a straightforward question? In this episode of Tech Talks Daily, I speak with Latané Conant, Chief Marketing Officer at Parloa, about the state of customer experience and what businesses must repair before agentic AI becomes another barrier between customers and support. Parloa's State of Agentic CX report assessed 10,000 enterprise websites, 4,000 chat interactions, and 100 phone trees. According to the company's findings, fewer than 10% of the tested chat conversations achieved the customer's goal. Only 1% of enterprises demonstrated readiness for automated agent-to-agent interactions. Those results raise a difficult question about years of customer experience investment. Businesses now have websites, chatbots, mobile applications, email, messaging, and phone systems, but customers frequently struggle to find help or complete the task that brought them there. Latané argues that part of the problem comes from treating customer service primarily as a cost center. When the objective is reducing contact volume, organizations can unintentionally make themselves harder to reach. This overlooks the commercial and operational information contained within customer conversations. Calls can reveal onboarding problems, unexpected product uses, recurring faults, and potential sales opportunities. Latané explains how analyzing service conversations can give marketing, product, operations, and executive teams a clearer picture of what customers are experiencing. We also examine why so many chatbots reproduce the frustration of traditional phone trees. Although the interface looks conversational, the system underneath may still rely on rigid categories and predefined routes. Customers then find themselves trying different words or repeatedly requesting a human agent. Latané describes a better agentic customer experience as being closer to talking with someone who already knows you. A personal AI agent could remember previous interactions, understand preferences, work across voice and text, and complete a request without making the customer repeat information. That possibility also introduces questions about trust, permissions, personal information, and oversight. Latané discusses the need to monitor what AI agents are doing, identify when conversations move away from approved subjects, and use supporting agents to detect potentially harmful behavior. Human involvement remains particularly important when a conversation involves distress, vulnerability, or emotional care. In Latané's roadside assistance example, AI can arrange a tow truck for a flat tire. If it detects signs that the caller is in distress, the conversation should move quickly to a person. We finish by considering what this means for customer service employees. Latané believes experienced representatives and operations teams can become builders and managers of AI agents, applying their customer knowledge across a much larger digital workforce. If the existing customer service front door is confusing and unwelcoming, should businesses repair that experience before inviting AI agents through it? Listen to the episode and share your thoughts with me.

What happens to tomorrow's leadership pipeline when employers automate the entry-level tasks through which beginners learn? In this episode of Tech Talks Daily, I speak with Gary Flowers, Chief Information Officer for Transformation and Technology Services at Year Up United, about AI fluency, human skills, economic mobility, skills-first hiring, and the future of entry-level work. Year Up United prepares young adults without bachelor's degrees for meaningful careers while helping employers reach skilled, career-ready talent. Gary says the organization has over 35,000 alumni working across companies ranging from the Fortune 1000 to the Fortune 50. Gary challenges the assumption that younger workers will automatically understand AI because they grew up with technology. Access to a chatbot does not create workplace readiness. Young adults also need training, support, ethical awareness, judgment, communication, adaptability, and experience applying tools to real business problems. He argues that AI may redefine entry-level work rather than eliminate it entirely. Candidates who understand how to work with AI may gain an advantage over those who do not, but employers must also reconsider which tasks beginners need to develop business knowledge and professional confidence. We discuss the risk of another technology divide. AI could widen economic opportunity, but unequal access to tools, training, mentorship, and workplace experience could reinforce existing inequalities. Gary believes organizations must teach workers when AI should be used, rather than limiting training to what the technology can do. Year Up United combines AI fluency with workplace and career readiness. Gary describes its 17 durable skills, six-month curriculum update cycle, close employer feedback loops, and participation as an inaugural host partner in Anthropic's Claude Corps Fellowship program. For employers, one of the hardest decisions involves balancing immediate efficiency with future capability. Automating junior work may reduce costs today while weakening the pipeline of experienced professionals and leaders required later. Gary recommends creating a culture of continuous learning, supplying employees with appropriate tools, building communities of practice, sharing successful use cases, and treating AI as a company-wide responsibility. He also distinguishes between AI as a workforce skill and AI as an organizational capability capable of changing how functions operate. Can employers capture the value of AI while preserving the career pathways that allow inexperienced workers to become tomorrow's experts and leaders? Listen to the episode and share your thoughts.

Can an app approved by Apple or Google still expose your business to security, privacy, and governance risks? In this episode of Tech Talks Daily, I welcome back Michael Covington, Vice President of Strategy at Jamf, for a conversation about the false confidence that can surround mobile security. Apple and Android provide strong protections, including app review processes, sandboxing, and device authenticity controls. However, Michael argues that a device being secure on day one does not mean it will remain secure throughout its working life. One of the most interesting points from our conversation is that mobile malware represents only a small part of the problem. Michael says it appears on fewer than 1% of the devices Jamf protects. The wider concerns include vulnerable third-party libraries, aging app versions, excessive permissions, unsafe web connections, compromised identities, software supply chains, and AI functionality introduced without the company fully understanding how it handles data. Michael also shares findings from Jamf analysis of corporate applications. According to the research he discusses, 95% of the apps examined contained at least one medium or higher severity vulnerability, 10% used vulnerable third-party libraries, and 96% included AI features. For security leaders, this creates a much broader question than whether an app contains malware. They need to understand what the app can access, where it communicates, how it handles data, and whether its capabilities comply with company policy. We discuss why familiar advice about updates, passwords, and suspicious links continues to fail when employees are busy or working from mobile devices on the front line. Michael explains how automation, clearer deadlines, and access policies can reduce risk without placing every responsibility on the user. The conversation also covers BYOD security and employee privacy. Modern Apple and Android controls can separate business information from personal apps, allowing employers to manage the work container without inventorying an employee's private digital life. Michael believes many organizations should reassess older BYOD programs that remain intrusive or unnecessarily restrictive. Finally, we examine the visibility security teams need across device configuration, patch levels, apps, permissions, identity services, web activity, and AI tools. Michael's advice is to start by understanding how people work before introducing heavier controls that may encourage workarounds and shadow IT. Does your organization know how its mobile risk changes after a device has been issued, or are you relying on the protection it had on day one? Listen to the conversation and share your thoughts with me.

How should finance leaders measure AI ROI when adoption has slowed and the cost of models, tokens and disconnected tools remains difficult to predict? In this episode of Tech Talks Daily, I welcome Jeremy Ung, Chief Technology Officer at BlackLine, back to the podcast to discuss how businesses can move from finance AI experimentation to operational deployment. Figures supplied for the interview show AI adoption in finance rising from 37% in 2023 to 58% in 2024, before moving only slightly to 59% in 2025. Jeremy argues that this apparent plateau reflects several pressures, including uncertainty around cost, regulatory requirements, auditability and the continuing debate over whether companies should build their own AI capabilities or purchase them through established platforms. Token spending is part of the problem. Unlike traditional software costs, model usage can be difficult to predict and allocate. Finance leaders want to understand whether applying AI to a workflow will produce enough value to justify that uncertainty. Jeremy believes companies should avoid creating artificial AI ROI metrics. The business measurements already exist. Is AI helping the company close its books faster? Is transaction matching becoming more accurate? Are collections improving? Is the work being completed faster or with fewer manual steps? We discuss what operationalizing AI in finance looks like in practice. Many processes still require employees to contact vendors, collect information, reconcile data and coordinate with other departments. Traditional software struggled with the variation found in these workflows, while AI can adapt to different processes and communication requirements. Accuracy and oversight remain necessary. Jeremy explains why companies need visibility into the prompts, reasoning, models, data, tools and permissions used by every AI agent. That information creates an operating record that finance teams, auditors and regulators can examine later. His analogy with food labeling provides a useful way to understand AI auditability. Consumers can inspect ingredients, calories and sourcing information before buying food. Finance leaders should expect comparable information about the models and data involved when an agent performs financial work. We also discuss the problem of fragmented data. Jeremy acknowledges the familiar rule of garbage in, garbage out, but argues that AI can help connect legacy platforms and mainframe systems that businesses previously found difficult to integrate. The role of finance professionals will change as agents perform additional work. Employees may spend less time completing individual tasks and more time setting goals, reviewing results, approving actions and directing teams of agents. Should CFOs continue buying additional AI tools, or concentrate on embedding existing investments into the financial workflows that determine business performance? Please share your thoughts with me.

What happens when an autonomous AI agent can complete thousands of actions before a traditional access review has even identified that something has gone wrong? In this episode of Tech Talks Daily, I speak with Alex Bovee, CEO and co-founder of C1, about identity security, runtime governance, shadow AI, and the controls companies need as humans and agents begin working together. Alex has spent much of his career in identity and security. He and his co-founder previously worked at Okta on zero trust products before creating C1 as an access control platform capable of operating at machine speed. That requirement has become increasingly important as AI agents begin accessing company data, calling tools, using credentials, and taking actions across enterprise systems. Alex describes agents as non-deterministic systems that can "reward-max." An agent may pursue its assigned objective so aggressively that it finds an unexpected or dangerous way to complete the task. It does not possess a moral compass or an intuitive understanding of what the company considers acceptable. Traditional identity processes were created for people. A company might review access every 90 days or investigate a security issue after an event. That approach becomes inadequate when an agent can execute thousands of actions within minutes. We discuss why identity is becoming a control plane for AI agents. Networks, data systems, and security tools all play important roles, but identity determines which resources an agent can access, which actions it can perform, and whether it acts independently or on behalf of a person. Without a defined identity or delegated authorization model, an organization may struggle to connect an agent's behavior with a responsible owner, a limited mission, and enforceable permissions. Alex explains the four connected capabilities inside C1's Agentic Control Plane. The first concerns shadow AI discovery. Companies need visibility across cloud services, SaaS applications, endpoint agents, hosted agents, local MCP servers, and credentials stored throughout the environment. This is particularly relevant because employees are downloading locally developed or "vibe-coded" MCP servers and running agent tools on their devices. These components can introduce software supply chain risks and expose local credentials. The second capability covers credential security. C1 has introduced a post-quantum credential vault designed to protect secrets and inject them into authorized agent workflows without leaving credentials scattered across devices and applications. The third area is runtime governance. Instead of reviewing behavior after an incident, organizations can evaluate an agent's actions against its assigned mission as they occur. If an agent is authorized to complete one business task but begins exploiting an internal tool, contacting an unapproved service, or attempting to extract data, runtime controls can block the action or request human approval. The fourth capability concerns agentic security intelligence. This uses information collected across identities, agents, permissions, credentials, and behavior to identify risks and support automated remediation. We also discuss human accountability. Alex says emerging regulatory thinking recognizes the need for a responsible person behind an autonomous agent. That connection allows businesses to establish ownership, delegate authority, and determine who remains accountable for the agent's behavior. The conversation then turns to the effect of AI on employees. Alex rejects the assumption that organizations will simply remove people as agents become more capable. His preferred analogy is that people are moving from manually producing every artifact to building and supervising the factory. Employees provide the inputs, direct the agents, examine the outputs, and correct the process when necessary. C1 has experienced this internally. Alex says its engineering team increased from roughly 150 weekly software merges to around 1,500, while engineering headcount grew by approximately 10% to 15%. That productivity requires careful human review. Generating work faster does not remove the need to assess whether the output is accurate, secure, useful, and aligned with the original objective. For CISOs and CIOs, the goal is to provide a governed path for AI adoption. A blanket prohibition may encourage employees to work around policy. Secure self-service access can give teams approved tools, defined permissions, and runtime protection. If an AI agent can operate at machine speed, are your organization's identity controls capable of observing, authorizing, and stopping it at the same pace? Listen to the conversation and share your thoughts with me.

What evidence would convince you that an AI agent is ready to make decisions involving employment, money, healthcare, or legal rights? In this episode of Tech Talks Daily, I speak with Vin Sharma, founder and CEO of Vijil, about the trust gap preventing many enterprise AI agents from progressing beyond proof of concept. Vin has spent approximately 30 years building software across security, operating systems, open source, cloud computing, machine learning, and AI. His previous work includes leading engineering at Amazon SageMaker and helping develop 11 AWS AI services. He argues that AI agents differ from conventional software because they combine autonomy with agency. They can interpret an objective, make decisions under ambiguous conditions, and take action. This raises a deeper question than whether an agent can complete a demonstration successfully: will it remain loyal to the interests of the person or business delegating the task? Trust is also specific to the job. Vin uses a simple analogy. You may trust a gardener to care for your lawn, but that does not automatically make the same person suitable to babysit your child. An AI agent must therefore be evaluated within the context of its users, task, operating conditions, authority, and potential consequences. Vin proposes testing three areas. Reliability asks whether the agent can perform its assigned task. Security examines whether it maintains its integrity when facing hostile or noisy conditions. Safety considers what happens when the agent fails and whether the resulting damage remains contained. This evaluation cannot end when the agent enters production. Models, integrations, data, users, and external conditions change. An agent may drift away from its original purpose, which means businesses need continuous monitoring, testing, and updating across the full AI agent lifecycle. We discuss how established security practices can be applied to this problem. Trusted execution environments, containment, least privilege, limited-duration access, and bounded models can reduce exposure. Smaller language models may also be better suited to narrow, high-risk tasks than a general model with broad permissions. Vin offers a three-part framework for governance: personas, purpose, and policy. Personas describe the people and attackers who may interact with the agent. Purpose defines the legitimate task. Policy sets the boundaries between permitted and prohibited behavior. For high-risk systems, his recommended starting position is that any action not explicitly permitted should be prohibited. A natural-language policy can then be converted into deterministic rules and controls governing the agent's behavior. Vin's most direct advice concerns evidence. Vibes, demonstrations, and benchmark scores do not prove that an agent is safe for a particular business process. A CISO should expect a complete risk assessment, while a business owner should receive proof that the agent will serve the organization's interests. His bridge analogy captures the issue perfectly. Engineers do not claim a bridge is safe because it looks impressive during a demonstration. They calculate load, tolerance, failure conditions, and provide test evidence. AI agents acting in consequential workflows deserve a comparable engineering discipline. If an agent developer asked you to trust their system today, would they be able to provide evidence of reliability, security, safety, loyalty, and contained failure? Listen to the episode and share your thoughts with me.

What if employees could access sensitive business applications from personal phones without storing company data on those devices? In this episode of Tech Talks Daily, I speak with Jared Shepard, CEO of Hypori, about virtual mobile infrastructure, BYOD security, employee privacy, zero trust, and the growing mobile threat created by AI. Jared's personal story deserves attention in its own right. He describes himself as a former homeless high school dropout who joined the Army, discovered an aptitude for IT, and applied what he learned to difficult technology problems in Iraq and Afghanistan. That experience gave him a firsthand understanding of what people working at the edge need from secure communications. The requirement that led to Hypori was unusually demanding. Users needed to obtain a phone from a local market, connect through a network assumed to be compromised, and access a protected enterprise environment without exposing sensitive information. Hypori's answer is virtual mobile infrastructure. According to the company, applications and enterprise data remain inside a protected cloud environment while the user receives a streamed visual experience. Sensitive data is not stored on the physical phone, tablet, or laptop. Jared explains why this differs from mobile device management. MDM attempts to secure, monitor, and control the endpoint. Hypori begins from the assumption that the endpoint may already be compromised. This can also protect employee privacy because the organization does not need visibility into the worker's personal device. We discuss how this approach could help government, defense, healthcare, banking, and smaller businesses that cannot maintain the same mobile security resources as a large enterprise. However, virtual infrastructure does not remove every responsibility. Organizations still need strong identity controls, protected cloud environments, reliable connectivity, policy enforcement, and careful vendor assessment. Jared also argues that AI is reducing the time between vulnerability discovery and exploitation. Security programs built around monthly patching may struggle when attack windows are measured in minutes. The conversation closes with leadership, resilience, and mentorship. Jared explains why hard work alone does not guarantee success and why valuable lessons can come from investors, generals, colleagues, friends, or the janitor who has spent 20 years observing how an organization works. Could virtual mobile infrastructure give employees secure access and personal privacy without forcing companies to control every device? Listen to the episode and share your thoughts.

Could your company be paying suppliers earlier than its competitors and unintentionally financing their advantage? In this episode of Tech Talks Daily, I welcome back Oliver Belin, co-founder and CEO of Calculum. Our previous conversation took place around ten years ago when Oliver was working with the Marco Polo Network and blockchain was attracting attention across trade finance. His latest venture concentrates on working capital, payment terms, and the role of AI in supplier negotiations. Oliver explains why working capital has moved higher on the agenda for procurement, treasury, and finance leaders. Companies can generate cash through sales, borrowing, inventory efficiency, faster customer collections, or changes to supplier payment terms. With borrowing costs higher and sales growth difficult in many markets, businesses are examining the cash already tied up within their operations. The difficulty is that procurement teams usually know their own supplier data but lack reliable information about the terms those suppliers accept from other customers. Negotiating without market benchmarks can lead to blunt policies, such as extending every supplier to 90 days. Oliver warns that indiscriminate extensions can create serious consequences. Smaller suppliers may experience cash flow pressure, increase their prices, reduce service, or direct capacity toward customers offering better terms. The buyer may improve its balance sheet while weakening an important part of its supply chain. Calculum uses transactional benchmark data to compare existing payment terms with the wider market. According to Oliver, the platform can show how frequently a supplier appears in its dataset, which terms it accepts elsewhere, and the probability that it will agree to a proposed change. AI and predictive analytics can then help companies concentrate on the suppliers where an adjustment would create the greatest financial impact and carry a higher probability of acceptance. This is particularly useful when an enterprise has tens of thousands of suppliers and procurement teams can only negotiate directly with a small proportion of them. Oliver says Calculum typically identifies free cash flow opportunities equivalent to approximately 8% to 11% of the spend analyzed. The amount identified does not automatically become realized cash. Procurement teams need targets, internal ownership, supplier conversations, and financing options to turn recommendations into results. He shares the example of an unnamed Fortune 500 pharmaceutical company that generated $227 million in free cash flow over 16 months. The program combined market-aligned payment terms with Supply Chain Finance, allowing participating suppliers to receive early payment in exchange for a discount based on the buyer's financial strength. Another UK company with approximately 4,000 suppliers generated €3 million in free cash flow within two months. Oliver attributes the speed partly to knowing which suppliers to approach first rather than attempting a broad, manual campaign. We also discuss supplier protection. Calculum identifies whether a business is a small or medium-sized enterprise, examines ultimate ownership, and considers financial strength. A financially vulnerable supplier may need early payment support rather than longer terms. Oliver's wider point is that AI cannot create reliable benchmarks from nothing. Useful predictions require traceable transactional data, clear objectives, and people prepared to act. Could better payment term intelligence improve your cash position while creating fairer, better-informed supplier relationships? Listen to the episode and share your thoughts with me.

What if the chemistry created by neurological disease could help activate medicine precisely where it is needed? In this episode of Tech Talks Daily, I speak with Sara Isbell, neuroscientist and co-founder of Enabled Therapeutics, about a proposed approach to one of medicine's most stubborn problems: delivering effective drugs to diseased brain tissue without exposing healthy areas to the same activity. Sara explains how the blood-brain barrier prevents many promising compounds from reaching the brain. When drugs do enter, they may spread across healthy and diseased regions alike, creating a difficult balance between therapeutic benefit and unwanted effects. We hear how an unexpected laboratory result led Sara and her co-founder to investigate whether pathological oxidative stress could convert a precursor molecule into a biologically active compound near the affected tissue. Sara describes this as pathology-gated therapeutic activation, where disease-associated chemistry provides the trigger that turns the medicine on. This remains developing science. At the time of recording, Enabled Therapeutics was preparing its first peer-reviewed manuscript and seeking partners to support further studies. Sara explains why reproducible evidence, regulatory guidance, and careful laboratory validation must determine whether the hypothesis advances. We also discuss how AI helps small biotechnology teams review literature, organize regulatory materials, connect ideas across scientific disciplines, and identify possible hypotheses. However, Sara offers an important reminder: AI can propose possibilities, but nature and experimental evidence decide what is true. Could following one unexpected result eventually offer researchers another way to approach neurological disease? Listen to the conversation and share your thoughts with me.

Could the real reason enterprise AI projects remain stuck in pilot mode be hidden inside the company's unstructured data? In this episode of Tech Talks Daily, I welcome back Oded Nagel, CEO of CTERA. We discuss why enterprise AI success depends on the condition, location, permissions, and business value of the data sitting underneath models and agents. Oded defines AI-ready data as information that is searchable, classified, and permission-aware. Many enterprises have petabytes of files distributed across offices, edge locations, legacy network-attached storage, and cloud platforms. Before introducing AI, leaders need to know what information they possess, where it resides, who can access it, and whether it remains valuable. The cost implications are significant. Copying every available file into an AI ecosystem can create expensive ingestion and storage bills. It may also reduce answer quality when stale, duplicated, irrelevant, or personal files enter the model's source material. Oded describes a customer classification project where approximately 80% of the data examined was stale or archival. The company also discovered personal content, including MP3 files, stored alongside enterprise information. Feeding such material into an AI system would consume resources without improving business results. We discuss Oded's recommendation to bring AI to governed data rather than moving data outside existing controls. Keeping intelligence close to the file system can preserve access permissions, audit logs, snapshots, and recovery mechanisms. Those protections become increasingly important when autonomous agents can read, move, modify, or delete files. Oded argues that every agent should be identifiable and its activity monitored. Businesses need to know which agent accessed which information, what action it performed, and whether the result can be reversed. Without those controls, a misunderstood instruction or malicious input could cause serious damage. The conversation also covers CTERA InsightAI, an agentic intelligence layer built into the company's data platform. Oded says it analyzes security activity and file-system metadata, allowing users to ask questions about stale data, file types, access patterns, deleted files, and ransomware impact using natural language. Rather than working through traditional dashboards and filters, users can question the data and request conclusions or recommended actions. Oded says some customers are piloting InsightAI while others already use it in production. For leaders measuring enterprise AI ROI, Oded recommends concentrating on storage costs, time savings, and speed to production. AI tools should make complex information easier to understand and reduce the time required to act. A ten-page report generated instantly provides limited value if nobody knows what decision to make from it. Does your company have enough visibility and control over its unstructured data to support production AI, or would classification uncover years of stale information and unnecessary expense? Listen to the conversation and share your thoughts with me.

What happens to the value of human judgment when AI makes execution faster, cheaper, and available to almost everyone? In this episode of Tech Talks Daily, I speak with Eric Wang, Vice President of Product and AI at QuillBot. Eric has worked in artificial intelligence since 2006, with previous leadership roles at Turnitin and Chegg. He now works on AI products used by millions of people to develop ideas, improve their writing, conduct research, and create new forms of content. Eric argues that AI's workplace impact extends far beyond automation. These tools are changing how people develop an argument, consider alternatives, cross traditional job boundaries, and turn an idea into something other people can understand. As technical execution becomes cheaper, Eric believes judgment, taste, and problem understanding become increasingly valuable. Someone with strong knowledge of a customer problem may be able to prototype software, produce marketing material, or develop a business proposal without depending on several specialist teams. That creates opportunities, although it also brings risks. AI can influence the direction of an argument, encourage misplaced confidence, and produce large volumes of content that sounds polished while saying very little. Eric shares an intriguing observation from QuillBot's user research: people increasingly refer to AI systems as "he" or "she." That small change in language may indicate that users are beginning to trust machines in ways they do not fully recognize. We also discuss how orchestrated workflows can give AI agents defined routes and boundaries, why Eric sees judgment and taste as durable business advantages, and what manual transmission cars can teach us about creativity in an automated world. Where should your organization draw the line between AI assistance and human judgment? I would love to hear where you stand, so will you share your thoughts with me?

Has enterprise AI finally reached the point where impressive demonstrations are no longer enough? In this episode of Tech Talks Daily, I speak with Bruce McMahon, Chief Product Officer at CallMiner, about what he describes as the industrialization of AI: the move from experimentation and excitement toward repeatable processes, measurable ROI, better customer experiences, and technology that can operate reliably at enterprise scale. Bruce explains why business leaders are increasingly asking a much simpler question about AI: how is this going to create value? Drawing on CallMiner's experience analyzing hundreds of thousands of hours of customer interactions every day, Bruce discusses how AI can surface operational inefficiencies and customer insights that were previously difficult to identify. The opportunity is not simply generating more data. Organizations need processes that get the right insight to the right person so something actually changes as a result. We also discuss how AI is changing workforce expectations. Bruce sees curiosity and adaptability becoming increasingly valuable, particularly among technical teams. As AI takes on more routine work, employees who question outputs, experiment with new approaches, and apply human judgment can become more valuable than those who rely solely on established technical knowledge. The economics of enterprise AI present another challenge. Foundation models, capabilities, and pricing continue to change rapidly, creating questions around vendor dependency and long-term costs. Bruce explains why companies may increasingly use a mixture of commercial, open-source, fine-tuned, self-hosted, and proprietary models rather than relying on one provider for everything. Governance becomes even more important as AI agents begin interacting directly with customers. We discuss red teaming, bias testing, compliance, data protection, monitoring, and why organizations need to decide which actions can be fully automated and which decisions must remain accountable to a human. Bruce also examines how AI is changing customer experience and the BPO industry. Rather than choosing between humans and AI agents, he sees value in designing systems where both can work together, with people handling interactions requiring judgment while AI manages high-volume and repetitive work. For CIOs, CTOs, COOs, customer experience leaders, and anyone responsible for enterprise AI strategy, this conversation provides a practical look at moving beyond AI pilots and turning the technology into a dependable part of business operations.  

What happens when an AI agent is compromised, manipulated, or simply does something nobody expected, but already has permission to access your most sensitive systems? In this episode of Tech Talks Daily, I speak with Geoffrey Mattson, CEO of SecureAuth, about why securing enterprise AI requires businesses to think beyond protecting models and start paying much closer attention to identity, authorization, access control, and what AI agents are actually allowed to do. Geoffrey argues that AI agents present a different security challenge from traditional software. Conventional applications can be tested against relatively predictable behavior. AI models are far less deterministic, particularly when prompt injection, excessive permissions, unexpected behavior, and autonomous actions enter the equation. His advice is to assume an agent could behave unpredictably and control what happens when it attempts to access a database, execute a financial transaction, call an API, or interact with another business system. We discuss what this means as companies race to introduce agentic AI. Geoffrey shares examples of employees granting AI tools permissions without fully understanding what they have approved, along with agents gathering information that creates unexpected privacy and compliance problems. This creates a difficult challenge for CIOs and CISOs. Boards want AI adoption because of its potential competitive value, while employees increasingly depend on AI tools to do their jobs. Simply blocking agents is unlikely to work. Security teams instead need mechanisms that allow innovation while controlling what those agents can access. Geoffrey explains why Zero Trust becomes particularly relevant here. Rather than authenticating a user or agent once and assuming it remains trustworthy, enterprises need to continually evaluate whether an action should be permitted at that specific moment. This leads to the concept of continuous authorization. Geoffrey explains how identity security is moving from asking "Who are you?" toward understanding intent, behavior, context, and authority for individual actions. This becomes increasingly important when one AI agent can create sub-agents, which can then create additional agents and pass permissions down the chain. We also discuss why agentic AI is exposing years of accumulated security debt. Many of the underlying problems are familiar: excessive privileges, inconsistent access controls, incomplete Zero Trust implementations, and systems that trust identities for too long. AI agents amplify those weaknesses because they can operate at machine speed. Geoffrey describes this as combining the unpredictability of humans with the power of machines. For CIOs, CISOs, security architects, identity teams, and business leaders deploying agentic AI, this conversation offers practical questions to ask before connecting agents to enterprise resources. What can the agent access? What authority does it have? Can that authority be reduced as tasks are delegated? Is every important action evaluated independently? And can access be revoked immediately when behavior changes? The goal is not to prevent organizations from using AI agents. It is to create a security layer that gives developers and employees room to experiment while ensuring agents only have the authority they need at the moment they need it. As autonomous AI becomes part of the enterprise workforce, identity alone may no longer be enough. Businesses increasingly need to understand intent, control authority, and continuously decide whether the next action should be allowed.

Companies are spending billions on GPUs, data centers, foundation models, and AI infrastructure. But what happens when the network connecting all of it cannot keep up? In this episode of Tech Talks Daily, I welcome back Avi Freedman, co-founder and CEO of Kentik, five years after our previous conversation. Avi has been operating large-scale networks since the 1990s, including more than a decade at Akamai, and brings a rare combination of founder experience and hands-on knowledge of how the internet actually works. We discuss why network performance is becoming an important factor in determining the return companies receive from their AI investments. If organizations cannot move data efficiently to models or deliver inference reliably to users and applications, expensive compute infrastructure can sit waiting while performance suffers and costs increase. Avi explains what technology leaders should measure to determine whether their network is helping or hindering AI workloads. This includes establishing performance baselines, synthetic testing across cloud and AI providers, understanding dependencies across the digital supply chain, and using observability to identify what changed when performance deteriorates. The conversation also examines network intelligence and why collecting telemetry alone is not enough. Organizations need to connect network data with the applications and users affected, understand historical behavior, determine which problems matter, and give network teams enough context to act quickly. Agentic AI introduces another opportunity. Avi explains how AI agents can increasingly perform the work of experienced network engineers by monitoring baselines, investigating alerts, troubleshooting problems, and recommending actions. But fully autonomous networks remain some distance away. Most enterprises currently want humans deciding whether significant production changes should be made. That leads us into governance. As businesses give AI systems access to increasingly important infrastructure, credentials, permissions, guardrails, and oversight become major considerations. Avi warns about ungoverned AI systems gaining proxy access to corporate infrastructure and explains why companies need clear boundaries around what agents can see and do. We also revisit a lesson from decades of internet infrastructure: individual components will fail. Rather than attempting to create networks that never fail, businesses should design for resilience through redundancy, over-provisioning, monitoring, and architectures capable of continuing when something inevitably breaks. For founders, CIOs, CTOs, network engineers, and infrastructure leaders building around AI, Avi offers practical advice on observability, network resilience, autonomous operations, AI infrastructure, and knowing when networking expertise should be developed internally or brought in from elsewhere. And we finish somewhere unexpected: how CEOs can use AI to make better decisions by explicitly asking it to disagree with them. Avi explains why turning AI from a sycophantic assistant into an argumentative colleague can expose weaknesses in an idea, improve communication, and help leaders test their thinking. AI may be transforming software, compute, and business operations, but none of it works without connectivity. As AI becomes part of the operational backbone of the enterprise, understanding the network underneath it becomes increasingly difficult to ignore.

Is AI really causing widespread job losses, or are a small number of announcements creating a much larger narrative? In this episode of Tech Talks Daily, I speak with Marvin Pohl, chief data scientist and cofounder of Clarecast, about AI layoffs, quiet restructuring, predictive workforce intelligence, and the responsibility that comes with forecasting company growth. Marvin's career began in physics and physical chemistry. After completing his PhD in Germany, he worked at Berkeley Lab and UC Berkeley before moving into data science at BASF. He describes how his role changed as generative AI entered the workplace. Initially, he encouraged skeptical colleagues to understand what language models could do. Today, he often finds himself warning people against accepting confident AI answers without checking the evidence. Clarecast was founded by Marvin, Jonathan, and CEO Bradley Taylor. The company combines employment profiles, job postings, technology adoption, stock information, industry data, and other signals to forecast how businesses may develop. Marvin says Clarecast covers over four million US companies and produces company-level forecasts extending 18 months. We discuss Clarecast's report on "quiet restructuring." The report considers whether AI-related workforce contraction may appear through slower hiring, unfilled positions, internal reorganization, automation, and the creation of new AI-related roles rather than widespread mass layoffs. Marvin says fewer than 100 companies in Clarecast's database had publicly attributed layoff announcements to AI. He describes this as a small proportion of the companies being analyzed and says projected US workforce growth appeared broadly flat rather than approaching a sudden collapse. However, Marvin is careful about what those findings can prove. The report presents a hypothesis, its model outputs are estimates, and correlation does not establish causation. Companies can change their hiring for many reasons, while employment data often takes time to reflect what has happened. Many of the AI-related announcements included in Clarecast's early analysis were also less than six months old. Marvin says a reliable assessment of whether companies followed through will require additional time because job postings, employment profiles, and reported headcount do not update immediately. We also discuss how Clarecast plans to apply its company intelligence to sales prospecting. Marvin argues that poorly personalized AI outreach is reducing response rates. Clarecast wants to help businesses identify a smaller number of companies that are showing signals of genuine need, allowing sales teams to spend additional time on relevant and personalized communication. How should business leaders use predictive intelligence without turning a probability into a predetermined outcome? Listen to the episode and share your thoughts with me.

What happens when an AI agent is authorized to make a payment, but nobody can verify the wider agreement behind it? In this episode of Tech Talks Daily, I speak with Zor Gorelov of Blue Language Labs about the infrastructure businesses may need as AI agents move from answering questions to negotiating, approving, purchasing, coordinating, and settling commercial activity. Many current business processes depend on human coordination. People reconcile spreadsheets, chase signatures, confirm deliveries, review exceptions, and resolve disagreements between systems. This work often remains invisible because employees absorb the ambiguity through emails, calls, and follow-up. Agent driven business changes the speed and volume of those interactions. One agent making an isolated payment can be handled as a software transaction. Several agents coordinating dependent actions across companies, banks, suppliers, platforms, and customers creates a much larger infrastructure problem. Zor argues that authorization answers only part of the question. An agent may have permission to pay, but every participant also needs to understand what the payment covers, which conditions apply, who can approve changes, what evidence confirms delivery, and when funds should be captured, refunded, or settled. Blue Language Labs is developing an open source protocol designed to structure those commitments. Blue Documents represent machine executable agreements containing participants, permissions, obligations, conditions, and the current state of a business process. Blue Mandates provide agents with revocable authority. A business can define spending limits, permitted actions, and thresholds requiring human approval. The meeting notes include the example of a restaurant operator allowing an agent to accept smaller bookings automatically while requiring approval for catering orders involving over 20 people. Blue Timelines provide an append only, hash linked record of actions, approvals, and changes. The aim is to give participants an independent history they can use when resolving disputes, instead of relying on conflicting emails or records controlled by one company. Zor brings the concept to life through a travel package assembled by an AI agent. The agent identifies a boutique hotel with spare inventory, a restaurant with available tables, and a local guide with unused capacity. Each business defines its terms, the agent assembles the offer, and the participants approve their roles. The customer purchases one package. Payment can be authorized at the beginning and captured according to agreed conditions as the hotel, restaurant, and guide confirm fulfillment. If one participant declines or fails to deliver, predefined rules determine whether the agent finds a replacement, changes the package, or triggers a cancellation. We also consider how Blue differs from traditional workflow systems, agent orchestration tools, and blockchain smart contracts. Blue is designed for coordination across separate businesses without requiring every participant to join one company platform or use global blockchain consensus. The opportunity could be especially valuable for smaller companies. Agents may allow several independent businesses to combine inventory, services, and expertise into offers they could not create individually. Adoption will depend on whether businesses, banks, and customers trust the protocol, accept shared definitions, and retain meaningful control. What would need to be written into a machine executable agreement before your organization could rely on another company's AI agent? Listen to the conversation and share your thoughts with me.

What separates an embedded finance partnership that changes customer behavior from an integration nobody would miss? In this episode of Tech Talks Daily, I speak with Rory Herriman, Chief Technology Officer and Chief Operations Officer for Zip's US business. Rory works across product, technology, operations, and business strategy, giving him a broad view of what happens after the API connection is complete and real customers begin using the service. Rory challenges a common understanding of embedded finance as placing one financial product inside another company's experience. Customers rarely wake up wanting embedded finance. They want to complete a purchase, manage their money, or solve a practical problem without an unnecessary interruption. The real test is whether the two businesses create something together that neither could provide independently. Rory calls this derived product market fit. Both products may succeed separately, but the combined experience must generate additional value for the customer if the partnership is going to last. Technology integration is only one part of the work. As businesses add customers and partners, they create new customer journeys, compliance obligations, servicing models, governance requirements, and operational processes. Rory argues that this complexity grows exponentially rather than linearly. This changes how technology teams should approach architecture. Instead of creating another custom integration for every opportunity, each partnership should contribute reusable capabilities to a wider platform. APIs, shared services, configuration tools, support processes, and governance models can then serve the growing ecosystem. We also discuss what partnership conversations reveal. Rory sees customer journey discussions during the first meeting as a positive sign. A conversation dominated by revenue division or integration mechanics may indicate that the participants have not established why the customer needs the combined service. His internal test is refreshingly simple. If the company launched the capability and removed it several months later, would the customer notice? If the answer is no, the partnership may have created technical activity without meaningful customer value. AI also enters the discussion. Rory believes AI can move financial services toward adaptive experiences where the product responds to the customer's circumstances. This offers opportunities for personalization and automated servicing, but it also increases the importance of responsible design, governance, customer consent, and clear accountability. For leaders building one-to-many embedded finance models, Rory's advice is to begin with the customer journey, establish alignment on values and service expectations, and build platforms that become stronger with each partnership. Would your customers miss the financial services you are embedding, or are they simply another feature occupying space in the journey? Listen to the episode and share your thoughts with me.

What happens to creator loyalty when somebody delivers the work, attracts an audience, and then waits weeks to be paid? In this episode of Tech Talks Daily, I speak with Rob Israch, President at Tipalti, about the payment infrastructure supporting the creator economy. Platforms may be able to add thousands of creators quickly, but the systems behind onboarding, tax collection, approvals, global payouts, communication, and reconciliation often struggle to keep pace. Rob cites research suggesting 87 percent of creators have experienced late payments. For a creator, payment is a direct test of whether a platform values their contribution. Delays, incorrect amounts, limited payment methods, or receiving funds in the wrong currency can damage trust and encourage successful creators to take their audiences elsewhere. This makes the payout experience part of creator retention. A platform may offer excellent creative tools and attractive commercial opportunities, but those benefits are easily undermined when creators have to chase payment updates or submit the same information repeatedly. Global growth adds another layer of difficulty. Rob explains that payment teams may need to account for approximately 26,000 rules, varying tax identification requirements, local payment methods, currency preferences, fraud checks, and screening against over five sanctions databases. If the correct information is not collected during onboarding, payment errors can increase by two or three times. The resulting problem concerns the complete workflow. Creator information must be collected securely, tax details validated, payment recipients screened, approvals completed, funds delivered through the preferred method, and every transaction reconciled with accounting systems. Creators also need timely communication when a payment is attempted, completed, delayed, or rejected. We discuss how automation can connect those stages and reduce the manual work that causes errors. Rob also describes practical roles for AI, including more responsive onboarding, automated fraud detection, tax validation, and immediate answers to payment-status questions. These capabilities can reduce support requests while giving finance teams more time to examine performance, risk, and growth. They also provide creators with something remarkably valuable: confidence that they will be paid accurately and kept informed when a problem occurs. Should creator payments remain a finance process, or should platforms treat them as part of the creator experience and retention strategy? Listen to the conversation and share your thoughts with me.

What happens when an enterprise AI agent can retrieve thousands of data points but cannot understand the customer, decision, or business moment in front of it? In this episode of Tech Talks Daily, I welcome back Boris Bialek, Vice President of Industries and Global Field CTO at MongoDB. We examine why the enterprise AI conversation has become more professional as organizations move beyond demonstrations and begin putting agentic systems into production. Boris argues that many companies do not have a shortage of data. Their problem is turning scattered data into information and then into usable knowledge. A bank balance is data. A complete view of a customer's relationship with the bank is information. Recognizing that the customer is currently researching a mortgage and may need assistance within the next 20 seconds is knowledge. This distinction leads to Boris's concept of a knowledge garden. Structured records, unstructured content, live signals, conversations, and business context are organized around a customer or outcome. Different departments can access the parts relevant to their work while AI agents receive the context needed to respond quickly. We also discuss integration debt. Boris recalls one system that required 18 seconds to assemble a customer view and says many enterprises are working with approximately 40 primary data sources. An agent can spend so much time coordinating access across APIs, caches, and applications that the business problem becomes secondary. Trust becomes equally important once an AI agent can act. Boris introduces two measures: the agent confidence score and the business risk score. The first evaluates whether an agent's output appears reliable based on its data, behavior, and context. The second considers the consequences of allowing that decision to proceed automatically. Together, these scores can help organizations decide which actions should pass automatically, which need further machine validation, and which should reach a human reviewer. Boris also explains why data lineage and complete audit trails must be designed into production systems from the beginning. For teams beginning this work, his advice is practical. Choose one business outcome, connect two or three relevant data sources, create a working prototype, and involve business and technical leaders in the same conversation. The goal is to demonstrate how data, context, confidence, risk, and human review work together before expanding the system. Does your organization have an AI data problem, or does it have a knowledge and context problem? Listen to the conversation and share your thoughts with me.

What happens when a warehouse management system believes stock is present, but nobody can find it on the warehouse floor? In this episode of Tech Talks Daily, I speak with Oana Jinga, co-founder of Dexory, who oversees the company's commercial strategy and product roadmap. Dexory has developed autonomous mobile robots capable of scanning inventory at heights of up to 18 meters while creating a continuously updated digital view of warehouse operations. The company says its robots have scanned one billion locations across 12 countries. Its customers include Maersk, DHL, Samsung, GE Appliances, Stellantis, GXO Logistics, and C.H. Robinson. However, the real story goes beyond the size of the robot or the number of locations scanned. It concerns what businesses can do once they have accurate information about their physical operations. Oana explains why warehouses often become data blind spots. Businesses usually know what entered the facility and what eventually left, but stock movements, damaged items, misplaced pallets, and inefficient use of space can remain difficult to track between those events. Dexory's robots scan approximately 10,000 to 12,000 pallet locations per hour. Oana recalls one customer discovering around £1.5 million in stock it had considered lost or written off. Other scans have revealed repeated pallet movements and potential opportunities to recover around 10% of warehouse capacity through better organization. We also discuss why visibility alone does not create business value. Dexory initially gave users large volumes of information, only to discover that extensive lists of problems could overwhelm warehouse teams. Its platform now prioritizes the actions requiring attention, helping users concentrate on a manageable number of issues each day. Oana explains why physical AI faces different challenges from software operating entirely within digital systems. Warehouses change constantly as people, vehicles, stock, temporary obstacles, damaged areas, and local working practices alter the environment. Robots and AI systems therefore require current physical data rather than relying on an old floor plan or assumptions recorded in another system. For companies considering warehouse robotics, Oana recommends starting with the operational problem. Leaders should observe how work happens, speak with employees about bottlenecks, define the result they want, and appoint an internal owner responsible for adoption. A robot sent to collect stock from an empty or incorrect location cannot complete its task, regardless of how capable its software may be. We also consider the future of warehouse work. Oana argues that robots can remove repetitive inventory walks and manual counting, allowing employees to interpret data, investigate problems, and improve operations. She also shares her experience as one of the few women in robotics a decade ago and explains why visible female role models can make the sector feel accessible to a wider group of people. Could physical AI help your warehouse teams make better decisions, or would inaccurate data and unclear ownership prevent the technology from delivering value? Listen to the episode and share your thoughts with me.

What can Formula One and football teach businesses about building customer relationships that continue long after a single event? In this episode of Tech Talks Daily, I speak with Ben Lewis, Vice President of Marketing at Infobip, about the company's work with AI-powered sports companions and what those experiences can teach customer experience leaders in every industry. Ben explains how Infobip worked with TGR Haas F1 Team to create RaceMate, an AI companion available through WhatsApp and Apple Messages for Business. Fans can access team information, driver histories, race schedules, trivia, personalized content and interactive experiences without downloading another application. We also discuss PitchMate, Infobip's conversational companion for global football fans. It remembers a fan's preferred team, can deliver personalized schedules and match information, and supports quizzes and other interactive features across an extended tournament. For me, one of the most useful lessons is the decision to meet fans inside messaging channels they already use. We have all downloaded an application for a conference, flight or one-off event, used it for several days and then forgotten it exists. RaceMate and PitchMate allow the conversation to remain available in the same place someone would message a friend. Ben also explains why Infobip measures success through returning users, conversation duration and the number of interactions rather than relying solely on clicks. TGR Haas F1 Team is currently using RaceMate to grow its fan community and provide useful content rather than constantly pushing merchandise. The same thinking can apply far beyond sports. We discuss travel companies contacting customers during unresolved claims, healthcare providers sending poorly timed automated messages and brands promoting products without recognizing that a customer is already involved in a dispute. Connected data can help prevent these disjointed experiences. Our conversation closes with practical advice for businesses adopting agentic AI. Ben recommends connecting customer data with campaigns, testing carefully, establishing guardrails and defining when an AI agent should transfer a conversation to a person. Are businesses investing too much in new customer applications when the better experience could already live inside WhatsApp, RCS or Apple Messages for Business? Please share your thoughts with me.

Late payments have become so common that many businesses simply accept them as part of commercial life. But should they? In this episode of Tech Talks Daily, I speak with Pat Bermingham, founder and CEO of Adflex, about why late payments continue to cost the UK economy an estimated £11 billion every year, why thousands of businesses fail because of cash flow pressures, and how technology could help change payment behavior rather than simply respond to it. Pat argues that late payments are rarely an administrative accident. In many industries they have become an informal financing mechanism, allowing larger organizations to protect their own cash flow while placing increasing financial pressure on smaller suppliers. Construction is one example, but the challenge extends across many sectors where long supply chains and uneven bargaining power make delayed payments the norm rather than the exception. We discuss why new government proposals to strengthen payment regulations represent progress, while also examining why legislation alone cannot solve a structural problem that has developed over decades. Instead, Pat believes technology can play a much bigger role. He explains how virtual commercial cards and Straight Through Processing (STP) allow buyers to access extended finance while suppliers receive payment far more quickly, without introducing additional friction into the payment process. Rather than forcing suppliers to accept card payments directly, the technology automates the process behind the scenes while improving reconciliation, increasing visibility and supporting healthier cash flow across the supply chain. The conversation also explores why many organizations still rely on fragmented payment systems created through years of acquisitions and disconnected technologies. Modernizing payment infrastructure can reduce delays, improve operational efficiency and help businesses build stronger supplier relationships rather than treating late payment as a normal business practice. Pat also shares how an earlier career as a music producer shaped his thinking about technology. Watching digital innovation transform music production helped him recognize how technology can simplify complex processes while also creating new business models that challenge established industries. For finance leaders, procurement teams, CIOs and business owners, this episode provides practical insights into improving cash flow, strengthening supplier relationships, modernizing payment processes and preparing for a future where prompt payment becomes both a commercial advantage and an increasing regulatory expectation. Changing payment legislation is important. Changing payment behavior is what will ultimately strengthen businesses, protect suppliers and create more resilient supply chains.

Most AI conversations begin with productivity. Joanna Pachnik thinks that's the wrong place to start. In this episode of Tech Talks Daily, I speak with Joanna Pachnik, founder of Blueclip, about why AI is changing far more than the speed of work. It's changing how businesses create value, what customers are willing to pay for, and what competitive advantage will look like over the next decade. Drawing on her experience leading global supply chain transformation projects at Ernst & Young and Mars before founding Blueclip, Joanna argues that knowledge is becoming increasingly accessible through AI. Research, analysis and reports that once took weeks and cost hundreds of thousands of dollars can now be produced in hours. That doesn't eliminate the need for expertise. It changes what expertise is worth. Rather than paying for information alone, organizations increasingly want implementation, measurable outcomes and practical experience that AI cannot easily replicate. Joanna explains why unique industry knowledge, benchmarking, practical experience and genuine human relationships may become more valuable as AI becomes more capable. We also discuss why so many enterprise AI initiatives struggle to deliver meaningful results. Joanna believes the technology is rarely the biggest obstacle. The real problem is poor data, undocumented processes and organizations trying to automate before building the foundations AI depends upon. Her advice is simple: prepare your data, document your processes, create a company knowledge layer, then introduce AI one use case at a time. The conversation also explores why AI should be viewed as a business transformation initiative rather than an automation project. Instead of accelerating existing processes, companies should ask whether those processes should exist at all. AI creates an opportunity to redesign how organizations operate, continuously improve decision-making and move people toward higher-value work. We also examine the importance of human oversight. Joanna believes AI should begin with people reviewing and guiding its outputs before gradually taking on more responsibility in carefully selected scenarios. Human accountability remains essential, particularly when AI supports material business decisions. For business leaders navigating AI strategy, digital transformation and enterprise innovation, this conversation offers practical advice on creating long-term value instead of chasing short-term AI hype. It explains why the companies that succeed will not necessarily be those using the most AI, but those prepared to rethink how they create value, organize knowledge and redesign their businesses around new possibilities. The future belongs to organizations that see AI as more than another productivity tool. It belongs to those willing to transform how they work, how they serve customers and how they create lasting business value.

What happens behind the scenes when you dial 911, and is the infrastructure ready for AI, satellite messaging, video, and precise location data? In this episode of Tech Talks Daily, I'm joined by John Snapp, VP of Technology at Intrado. John has spent around 23 years working with cellular, location, and 911 technologies. He explains how a mobile emergency call is located, routed through a dedicated network, and directed to the appropriate Public Safety Answering Point. We discuss where AI can provide practical support inside emergency communications. Translation can help telecommunicators understand callers without waiting for an interpreter. Real-time transcription can capture details and suggest established procedures. AI voice agents can also handle suitable nonemergency inquiries, giving trained staff additional time for calls where lives may be at risk. John is clear that emotional emergency calls still demand human understanding and authority. AI can supply information, identify possible synthetic voices, and reduce administrative work, but trained telecommunicators remain responsible for interpreting the situation and directing the response. Our conversation also examines the infrastructure beneath these capabilities. Legacy 911 networks were designed largely for voice and limited amounts of data. Next Generation 911 introduces IP connectivity capable of carrying text, images, video, and richer location information. John explains how this foundation has made satellite texting possible and why similar capabilities were far slower to introduce using older networks. Moving to NG911 creates its own problems. Different vendors can comply with the same technical standard while implementing it differently. Calls may also need to move between modern and legacy call centers, making interoperability testing between jurisdictions a major part of deployment. We also consider cloud resilience, local survivability, connectivity diversity, telephony denial of service attacks, AI generated swatting calls, and the danger of adopting automation before establishing governance. John recommends starting with lower-risk areas such as quality assurance and nonemergency calls, communicating openly about AI use, and expanding only after teams understand the operational impact. As emergency communications become richer and increasingly connected, how should public safety agencies balance faster innovation with the reliability and human judgment every caller depends on? Listen to the episode and share your thoughts with me.

When organizations review their cybersecurity posture, printers are rarely the first systems that come to mind. Yet they often account for around 20% of network endpoints while receiving, storing, processing, and transmitting sensitive business information every day. In this episode of Tech Talks Daily, I welcome back Jim LaRoe, CEO of Symphion, to discuss why printers and other connected IoT devices have become one of the most overlooked areas of enterprise cybersecurity and why AI-powered attacks are raising the stakes for organizations that continue to ignore them. Jim explains how many businesses continue to treat printers as simple office equipment rather than Linux-based network devices with privileged access to email systems, file servers, identity services, and critical business workflows. Because responsibility for these devices often sits between procurement, managed print providers, IT operations, and security teams, they can easily fall outside normal cybersecurity processes. We discuss how the threat landscape has changed over the past year as AI enables attackers to automate reconnaissance, credential theft, lateral movement, and ransomware deployment. Jim explains why organizations adopting Zero Trust principles also need to rethink how they secure and manage connected endpoints that have traditionally been overlooked. The conversation also explores certificate lifecycle management, cyber hygiene, firmware management, endpoint visibility, and why unsupported devices can introduce unnecessary risk into modern enterprise environments. For organizations managing hundreds or even thousands of printers across multiple locations, Jim explains why protecting these endpoints does not need to create additional operational burden. Instead, security should become an ongoing operational program that continuously monitors devices, detects configuration drift, applies security controls, and helps organizations maintain compliance without disrupting critical business workflows. We also discuss the governance challenge many organizations face. Before companies can reduce risk, someone needs to own it. That means establishing accountability, assigning budget, understanding which devices exist across the business, and recognizing that printers and IoT devices deserve the same attention as servers, laptops, and other managed endpoints. If you're responsible for cybersecurity, IT infrastructure, risk management, or digital transformation, this episode offers practical advice on protecting forgotten endpoints, strengthening Zero Trust strategies, improving endpoint visibility, and reducing the hidden risks that AI-powered attackers are increasingly looking to exploit. Sometimes the biggest cybersecurity vulnerability isn't the system you forgot to patch. It's the one you forgot was connected in the first place.

What happens to digital asset ownership when the cryptography proving that ownership can no longer be trusted? In this episode of Tech Talks Daily, I speak with Yoon Auh, cofounder of BOLTS Technologies, about quantum computing, blockchain security, and the need for crypto agility. Yoon brings an unusual perspective to the subject. Before moving into applied cryptography, he spent years building and operating high performance trading systems at firms including Credit Suisse, Goldman Sachs, Geode Capital, and Magnetar Capital. Yoon explains that blockchain ownership ultimately depends on digital signatures and public keys. Most major blockchain systems use variants of elliptic curve cryptography because it has historically offered speed, compact signatures, and dependable protection. However, sufficiently powerful quantum computers could eventually challenge the mathematics supporting that protection. The risk does not begin when such a quantum computer arrives. Yoon describes how attackers can collect encrypted traffic today, store it, and attempt to decrypt it later. This creates an immediate concern for governments, financial institutions, and businesses holding information that must remain private for many years. We also discuss QFlex, the post quantum ready API developed by BOLTS Technologies. The company describes its approach as cryptographic logistics, allowing different cryptographic methods to be selected at the transaction level. Yoon argues that a small payment and a multimillion dollar asset transfer should not automatically receive identical protection, particularly when stronger cryptography may require additional processing, storage, and cost. Another concern is uncertainty around the available post quantum algorithms. Yoon explains that cryptographic methods can survive years of examination before a weakness is discovered. His argument is that organizations need the ability to change algorithms quickly if one becomes vulnerable, rather than making a permanent choice and hoping it survives every new attack. The conversation also examines digital asset sovereignty. Who decides how a transaction is protected: the platform, the protocol, or the asset holder? BOLTS Technologies believes that choice should return to the holder, while QFlex aims to provide that control without hard forks, network downtime, or protocol changes. The interview also covers the company's research background and its pilot work with the Canton Foundation. Yoon closes with a lesson from his trading career. Backup and failover exercises often failed because they were treated as occasional events. His advice is to make exceptional processes routine, ensuring that the organization has already practiced changing systems before the moment arrives when it has no other option. Should digital asset holders control the cryptography protecting every transaction, or should platforms continue making that decision for them? Listen to the episode and share your thoughts with me.

What prevents a successful AI experiment from becoming a dependable production system that delivers measurable business value? In this episode of Tech Talks Daily, I speak with Ed Macosky, Chief Product and Technology Officer at Boomi, about AI pilot purgatory, integration, governance, model selection, token costs, and the technical skills businesses may need as adoption grows. Ed leads Boomi's product and engineering teams while also using AI tools inside his own organization. That gives him a view from both sides: creating technology for enterprise customers and applying it within active product development workflows. He believes many AI pilots begin with the wrong question. Teams become interested in the latest model or feature before defining the business problem they want to solve. The experiment may work during a demonstration, then fail when it encounters real data, access controls, security policies, and production systems. Placing company information inside a data lake and adding a language model does not automatically create a business application. The system must access current data reliably, respect employee permissions, connect with existing applications, and operate within governance rules that security teams can approve. Ed recommends beginning with a defined business opportunity and establishing the access required to support it. Existing APIs can already provide authentication, permissions, and governance. MCP can offer another route into enterprise systems, but those connections still require security, monitoring, and management. Team alignment also matters. An AI center may be racing to test models while an integration center concentrates on a different set of priorities. When those groups fail to coordinate, the pilot lacks the connectivity and automation required to become part of a production workflow. The discussion then turns toward fragmentation. Every technology wave produces new vendors, frameworks, and specialist tools. Early experimentation benefits from variety, but mature companies can eventually find themselves maintaining a complicated collection of products held together with custom code and, occasionally, the digital equivalent of duct tape. Ed does not recommend placing every function with one provider. He does argue for enough consolidation and abstraction to prevent experimentation from creating years of technology debt. Governance and observability should also work horizontally across different environments, including platforms such as SAP, Salesforce, and several AI model providers. That becomes increasingly important as businesses introduce autonomous agents. Leaders need to know which agents exist, what systems they can access, what actions they can take, and how each decision is recorded. AI gateways and agent control towers can provide a wider view across otherwise separate technology environments. Ed also introduces the idea of the frontier engineer. A prompt engineer concentrates on communicating effectively with a model. A frontier engineer understands how the model works, including its logic, mathematics, algorithms, and suitability for different workloads. He does not believe every company needs a large team of these specialists. However, he argues that enterprises need at least one person capable of assessing vendor claims and deciding whether a frontier model, specialist model, or open weight model fits a particular workload. Cost creates another reason to examine model selection. Sending every employee request or agent task to the most capable frontier model can become expensive. Some repeatable workloads may run on open weight models inside the company's cloud or hardware environment, giving finance teams greater cost certainty. Boomi is developing Boomi Prompt to route requests according to their complexity and requirements. A simple factual request might go directly to an API. A forecasting task may use a smaller model. A difficult analytical request could be sent to a frontier model. Ed uses the weather as a helpful example. Retrieving next Tuesday's forecast does not require a language model when a public weather API can return the answer directly. Asking a model to perform every form of automation wastes tokens, computing power, energy, and money. The episode closes with practical advice for CIOs. Avoid starting with a broad objective such as agentifying the entire business. Choose a department, identify a small number of tasks, define the expected return, and work backward. Once the team proves value and understands the operating requirements, it can repeat the process elsewhere. Could intelligent routing, stronger integration, and clearer business outcomes finally move enterprise AI beyond pilot purgatory? Listen to the episode and share your thoughts with me.

What must happen before a business can trust AI agents to detect and resolve operational problems without waiting for human intervention? In this episode of Tech Talks Daily, I speak with Josh Clay, Regional Vice President of Solution Engineering for Dynatrace in the UK, about autonomous operations, AI observability, fragmented telemetry, business outcomes, and the growing pressure to control token and data costs. Josh has spent much of his 11 years at Dynatrace discussing the road toward autonomous operations. The earliest version involved reducing the time organizations spent inside IT war rooms. He remembers calls with 30 or 40 people attempting to establish which team was responsible for an incident. He jokingly calls this the "mean time to innocence." Modern observability reduced many of those investigations from several hours to between 30 and 60 minutes. Agentic AI creates the possibility of going further by identifying a problem, understanding its cause, and resolving it before the customer experience is affected. That ambition also introduces risk. Josh cites Dynatrace research showing that 52% of respondents view security, privacy, and compliance concerns as barriers to AI adoption. He believes many organizations still lack full observability across their existing technology environments, making autonomous agents harder to supervise. Josh shares a warning from Alex Hibbert of Storia Group: AI can amplify existing technology problems. If telemetry is fragmented, data quality is poor, or teams cannot see how services depend on one another, adding autonomous agents may increase the speed and scale of the resulting failure. Trust therefore depends on visibility. Josh describes observability as a control plane for agentic AI because it can show what an agent is doing, why it made a decision, and what happened afterward. Defined guardrails and real time information can give leaders confidence without asking them to surrender control blindly. The adoption figures show how early this work remains. Josh says 50% of businesses have AI operating in limited production use cases, often performing one isolated task. Only 23% describe their deployments as connected across the wider organization. We discuss how observability has progressed beyond technical monitoring. An airport can measure whether technology changes improve e-gate availability and passenger processing times. A bank can examine whether application performance affects mortgage completion rates. These connections allow leaders to measure AI through business results rather than relying entirely on response times and infrastructure metrics. Reliable agents also need suitable data. Dynatrace says AI agents operating with deterministic data can work 12 times more accurately and three times faster while using two and a half times fewer tokens. These remain company findings, but they demonstrate why context and causality can affect cost as well as reliability. Fragmented telemetry creates another barrier. Logs may sit in one platform, front end monitoring in another, and metrics or traces somewhere else. Attempting to reconstruct every relationship for an AI agent can become expensive and difficult. Josh recommends bringing observability data into a connected environment where relationships between services, cloud resources, traces, metrics, and logs are already understood. He also warns against collecting information simply because it exists. Data hoarding increases ingestion costs and can introduce personal information into systems without a clear business need. The conversation then moves toward AI FinOps. Leaders want to know what agents cost, how many tokens they consume, and whether those costs produce a measurable return. Josh describes a Dynatrace proof of concept that identified potential annual savings just below £250,000 within one small environment. That example reinforces a recurring concern. Organizations are racing to place AI into production, then moving to the next project without reviewing whether the previous environment is appropriately sized or financially efficient. Josh hopes companies will develop a more pragmatic view of AI as another enterprise tool. That means establishing agreed methods for deployment, monitoring, cost management, incident response, and measuring business results. Could observability provide the confidence businesses need to move from isolated AI experiments toward autonomous operations? Listen to the episode and share your thoughts with me.

How can businesses secure AI agents that read sensitive information, update systems and communicate with other agents on behalf of employees? In this episode of Tech Talks Daily, I speak with Sachin Nayyar, founder and CEO of Saviynt, about AI agent identity security and the controls businesses need before autonomous systems enter production. Saviynt manages over 100 million identities for over 700 customers. Sachin explains how enterprise identity has expanded beyond employees to include partners, applications, machines and autonomous AI agents. An AI agent creates a different access problem because it is both an identity and an application. It can receive permissions, access information and perform actions, but its behavior can also be governed through software while it is being developed and while it is operating. Sachin uses an HR copilot to demonstrate why context matters. Two employees can ask the same question about salaries but should receive different answers based on their roles, locations and applicable policies. Those decisions must be evaluated while the request is being processed without creating delays that make the system unusable. The risk grows when an agent crosses from one technology environment into another. An agent built within Microsoft may need to access Salesforce, ServiceNow, Jira or another external system. Sachin warns businesses never to solve this problem by giving an agent a permanent administrative account. We discuss Zuma, Saviynt's identity security platform for AI agents and non-human identities. Sachin describes a four-part framework beginning with agent discovery and a central registry. Every agent should then receive one accountable human owner, temporary access for its assigned task and policy enforcement while it acts. Ownership becomes especially important when an employee leaves. Saviynt's approach begins an automated reassignment process and blocks actions if an agent attempts to operate without a current owner. The relevant security team can then investigate before allowing further activity. Sachin also explains why identity controls should enter the development process rather than being added after deployment. Saviynt is working with LangChain and other agent development platforms to make identity policies available while AI agents are being built. The conversation also covers Saviynt's partnership with Zscaler. Zscaler provides inline enforcement, while Saviynt contributes identity information about the agent, its owner, existing permissions and expected behavior. Sachin closes with an optimistic argument. Because businesses can place security controls into the code and enforce them while agents act, AI workloads may eventually become better governed than traditional human access. Could every AI agent inside your business be traced to one accountable owner, one approved purpose and a limited set of temporary permissions? Please share your thoughts with me.

What becomes possible when enterprise computer vision no longer depends on expensive GPU infrastructure? In this episode of Tech Talks Daily, I speak with Glenn Jocher, founder and CEO of Ultralytics, about YOLO26, CPU inference, edge AI, open vocabulary vision, deployment economics, and the practical work required to move computer vision from a promising pilot into production. Glenn's route into AI began inside the U.S. intelligence community. He worked with the National Geospatial Intelligence Agency and Defense Intelligence Agency on particle physics applications, attempting to detect and track antineutrinos. Antineutrinos are extraordinarily difficult to detect because they pass through almost everything. Glenn describes them as the perfect spy. While searching for better detection methods, he discovered that computer vision researchers were solving similar problems with images. His original attempt to transfer those techniques into particle physics did not succeed. However, the work introduced him to a field where the technology could create a visible effect on everyday life. That led him toward open source development and eventually the YOLO models for object detection, classification, segmentation, and tracking. Glenn believes computer vision research has historically placed too much attention on small gains in accuracy while overlooking deployment economics. A model can perform impressively inside a laboratory and still remain unsuitable for a factory, warehouse, store, vehicle, drone, or medical environment. Price, latency, power consumption, data privacy, and deployment speed can determine whether the technology is commercially useful. This led Glenn and Ultralytics toward smaller models capable of running close to where images and video are generated. YOLO26 continues that approach with architectural changes designed specifically for CPU inference. Glenn says the model can process camera streams in real time at 30 frames per second and run across Intel CPUs, AMD CPUs, and lower power devices such as Raspberry Pi computers. This matters because specialist GPUs can increase the equipment cost and power requirements of a computer vision project. Running inference on existing CPUs or edge hardware can make deployment economically possible across larger numbers of cameras and locations. The scale already involved is difficult to comprehend. Glenn says Ultralytics models now process approximately three billion inference jobs each day, equivalent to around 30,000 every second. These jobs include images, videos, and collections of images being analyzed to detect, segment, or track objects. He attributes the platform's maturity to thousands of mistakes and bugs corrected through a rapid feedback cycle. New models are released, users report problems and request features, and the team incorporates that information into later versions. We also discuss the respective roles of cloud and edge infrastructure. Glenn sees cloud platforms continuing to provide the computing power required for training, while computer vision inference often belongs at the edge. Local processing can reduce latency, control operating costs, and keep sensitive video or medical information closer to where it was created. The smallest YOLO model is approximately three megabytes, according to Glenn. That allows it to reach mobile phones, vehicles, drones, battery powered devices, and other environments where a large language model would be impractical. Open vocabulary vision provides another development. Traditional object detection models are trained to recognize a fixed collection of objects. If a model learns to detect dogs and the user later wants it to detect cats, retraining can cause it to forget earlier knowledge unless both categories appear in the new training data. Glenn explains how promptable models can identify common everyday objects from text or visual instructions without additional training. A user could request a person wearing a blue shirt and white shoes, for example, and the system could search an image for that description. That flexibility could benefit businesses whose requirements change regularly. It reduces the need to create and label a new data set every time the company wants the model to recognize another common object. The range of current applications is already extensive. Glenn describes YOLO being used across robotics, parking, industrial safety, PPE detection, warehouses, aviation, security, traffic management, food quality, and manufacturing. Some of his favorite examples involve environmental problems. One company uses YOLO with underwater vehicles to identify and recover plastic from the ocean. Other applications detect smoke and fire early enough to support forest fire response. For leaders considering computer vision, Glenn recommends beginning with a defined problem and measurable outcome. A manufacturing company may want to reduce defects, but it still needs labeled examples showing the model what acceptable and defective products look like. He advises testing the idea through a limited pilot, measuring the return, and expanding only when the evidence supports further investment. Computer vision has become easier to deploy, but practical problems involving data, cameras, integration, reliability, and operating conditions still separate a demonstration from a production system. Could CPU inference and open vocabulary models make computer vision practical for processes your organization previously considered too expensive? Listen to the episode and share your thoughts with me. Useful Links   Ultralytics website Ultralytics Platform      

What can private companies learn from government caseworkers about adopting automation and using data more effectively? In this episode of Tech Talks Daily, I speak with David Turner, General Manager and Senior Vice President of Government Services at Equifax Workforce Solutions, about public sector automation, data modernization, and the people responsible for delivering social services. The conversation begins with a surprising finding from an Equifax Government Services study of over 500 U.S. government employees. Every respondent expected efficiency to improve during the following year, while 95% believed automation would free time for higher value, human centered work. David explains why government employees may be more receptive to modernization than many people assume. Caseworkers operate under rising demand, staffing pressures, changing policy, and limited budgets. When technology removes repetitive administration or supplies information faster, they can see an immediate connection between the tool and the person waiting for support. We discuss how the meaning of automation has changed inside social services. A few years ago, it might have meant entering information into an online portal. Today, integrated connections can search data sources behind the scenes and return verified information during the caseworker's existing process. David describes continuous evaluation, which can identify when circumstances within a caseload have changed. Instead of searching every case for a possible update, a worker can direct attention toward the people whose income, address, or employment data indicates that further review may be needed. Income verification provides another example. Equifax says it can return income information in under one second, helping prevent the delays created when an applicant must leave the process to find a document. Those pauses matter when an eligibility decision already involves several stages and numerous external systems. The gig economy makes this work harder. Applicants may receive income from employment, contract work, digital platforms, and several side projects. Agencies need access to a fuller income picture without sending people back toward paper forms and manual verification. David also shares what Equifax has learned through its Day in the Life program. The team spends time with caseworkers to understand their processes, policy restrictions, and information constraints. He believes public sector leadership can remain closely connected to employees in the field because many agency executives previously performed those roles themselves. AI introduces further possibilities. David discusses how data and AI could eventually identify signs that someone who has left an assistance program may be experiencing financial difficulty again. Community groups, food banks, or other services could potentially offer support before that person returns to crisis. Such a system would also require careful decisions around consent, privacy, accuracy, and responsibility. The central lesson is refreshingly human. Technology creates value when it removes administrative pauses and gives experienced caseworkers additional time to understand someone's circumstances. Could public sector automation teach private companies how to connect technology investment with human outcomes? Listen to the episode and share your thoughts with me.

What happens when an AI experiment becomes a production service that your employees, customers, and daily operations depend upon? In this episode of Tech Talks Daily, I speak with Brian Klingbeil, Chief Strategy Officer at Ensono, about AI infrastructure resilience, operational dependency, FinOps, legacy modernization, and the growing pressure to prove that enterprise AI investments are producing meaningful returns. Brian has been speaking with major enterprises through Ensono's Executive Advisory Council. Three years ago, many participants were experimenting with proofs of concept. Today, they are being asked to present AI projects that are already in production, approaching production, or demonstrating a clear return through productivity, lower risk, service quality, or financial results. That progression creates a new problem. When an AI model begins supporting product delivery, customer service, logistics, software development, or internal operations, it becomes part of the company's operating infrastructure. Leaders must then ask familiar IT questions about availability, monitoring, security, incident response, disaster recovery, ownership, and cost. Brian believes FinOps often provides the first warning. Token consumption can be difficult for CFOs and business leaders to interpret, particularly when hundreds of agents are operating across different models. Ensono's internal platform has produced around 1,000 agents, prompting questions about which are effective, which are expensive, and who should carry the cost. We discuss why chargeback and showback could change employee behavior. When AI spending is absorbed by a central corporate budget, teams may have little reason to question whether an expensive model is suitable for a routine task. When the cost reaches their departmental budget, the decision can look very different. Architecture also matters. Brian recommends systems that are loosely coupled and tightly integrated. Companies should be able to replace a model, provider, FinOps tool, or service as the market changes, while still connecting each component closely enough to deliver useful business outcomes. That creates a genuine tradeoff. Providers such as Microsoft, Amazon, Google, OpenAI, and Anthropic can offer specialist capabilities that businesses may want to use. Avoiding every provider specific feature can limit what the technology delivers, while becoming too dependent on one provider can make future change expensive and disruptive. The conversation then turns toward legacy technology. Brian argues that many systems described as outdated still process airline reservations, banking transactions, insurance claims, government services, and other high volume workloads. Turning them off without suitable replacements would create far bigger problems than the word "legacy" suggests. AI can change the modernization decision. Ensono worked with Markerstudy Group to analyze six million lines of RPG code running on an IBM i platform. The resulting plan identified applications that should move elsewhere while preserving workloads that still benefited from the platform's reliability and transaction processing capabilities. Brian treats migration as one possible part of modernization. AI tools can document old code, support modern development environments, and allow younger developers to work with established platforms without immediately beginning a lengthy and expensive replacement program. We also discuss Ensono's use of AI operations. Brian says the company reduced mean time to repair by 50% while processing approximately 50,000 tickets each month. The example shows how AI value can be measured through service quality and operational performance rather than relying entirely on direct revenue. The result is a balanced conversation about moving quickly while building enough control to keep AI dependable. Organizations need space for experimentation, but production services also require ownership, budgets, recovery planning, and people who know what to do when something fails. If one AI model or provider disappeared tomorrow, how much of your business would stop working? Listen to the episode and share your thoughts with me.

What happens when an autonomous AI agent makes a financial decision using inaccurate, outdated, or poorly synchronized data? In this episode of Tech Talks Daily, I speak with Marcin Kaźmierczak, cofounder of RedStone Oracles and Credora Ratings, about why verifiable data is becoming so important to financial AI agents. RedStone originally developed its oracle infrastructure to supply smart contracts with reliable information from hundreds of sources. The same principles are now being applied as AI agents begin analyzing markets, recommending allocations, processing payments, and executing trades. Marcin explains what a blockchain oracle does and why smart contracts cannot independently access real world information. RedStone aggregates, cleans, and distributes financial data, including asset prices, liquidity, volatility, and market capitalization. According to Marcin, its network currently secures over $10 billion in total value locked and has operated for six years without a mispricing or downtime event. The discussion then moves into agent driven finance. AI agents can process information and act at considerable speed, but that speed introduces problems when the underlying information is delayed or the model hallucinates. Marcin describes the synchronization challenge created when an oracle updates every five seconds while an agent makes decisions every second. One example captures the risk. An AI trading agent reportedly generated $200,000 over three months before losing $250,000 in two transactions. Marcin explains how Credora Ratings can add financial risk context by rating assets and strategies from D to A. Companies can then instruct an agent to operate only within an approved risk range. We also discuss tokenized assets, the growing interest from major financial institutions, and why blockchain networks offer an attractive operating environment for autonomous finance. Marcin shares practical advice for leaders, including speaking with experienced implementers, testing agents in closed environments, identifying likely failure scenarios, and creating response policies before introducing real money. What evidence would you require before trusting an AI agent with a financial decision? Listen to the episode and share your answer with me.

What if a restaurant could identify a margin problem while the ingredients were still being unloaded, rather than discovering it several weeks later? In this episode of Tech Talks Daily, I speak with David Cantu, CEO of Craftable, about AI restaurant back-office technology, margin intelligence, inventory management, purchasing, invoice automation, and the continuing importance of human hospitality. David has spent decades working in restaurants and technology. He describes an industry dealing with staffing difficulties, rising leases, food inflation, lower traffic, and relentless margin pressure. David cites a National Restaurant Association study indicating that 40% of restaurateurs were not profitable during the previous year. Craftable connects purchasing, recipe management, inventory, accounting, sales data, and analytics. The company says its platform is used by over 10,000 restaurants, hotels, and venues. David argues that useful hospitality AI should automate the work that keeps managers, chefs, and operators away from guests. This includes producing sales forecasts, suggesting orders, planning preparation, identifying invoice anomalies, and comparing projected labor with actual requirements. The conversation becomes particularly practical when David describes a restaurant receiving a ribeye that has increased in price by 20%. If the dish is a popular, high-margin menu item, that vendor increase can quickly reduce its profitability. Craftable's Invoice AI can detect the change when the invoice is received. The operator can then consider running a higher-priced chef's special, promoting another steak, reviewing the menu price, or adjusting future orders. Waiting until month-end reconciliation would explain the lost margin but leave no opportunity to recover it during service. We also discuss the difference between AI intelligence and operator knowledge. AI can examine large volumes of information, but an experienced restaurateur understands the atmosphere, the team, the guests, and what is happening at that particular moment. David believes AI recommendations need to show their work. Managers should be able to inspect how a sales forecast, suggested order, staffing plan, or trend was calculated. Transparency helps people assess the recommendation without forcing them to search through another large analytics report. Craftable is also testing how to measure whether a recommended action produced a result. If a manager coaches a server with unusually high complimentary items or promotes a menu category with falling attachment rates, the platform can examine whether that action affected sales or margins. David is skeptical of hospitality AI added as a promotional layer without being built into the daily workflow. At industry conferences, he has seen vendors attach language models to existing products while offering little operational value. His hope for restaurant AI is highly human. He wants technology reducing administrative pressure while employees welcome guests, serve meals, develop their teams, and create the warm experiences that define hospitality. Which restaurant decision could protect profitability if the operator received the right information before the next service began? Listen to the episode and share your thoughts with me.

Why do so many enterprise AI initiatives begin with impressive demonstrations but struggle to produce measurable business value? In this episode of Tech Talks Daily, I speak with Dom Selvon, CTO and value partner at Valiance, about enterprise AI ROI, outcome-based consulting, build versus buy decisions, proprietary data, ontologies, and governance. Valiance is an AI-native consultancy that charges against client outcomes rather than hours worked. Dom explains why his "value partner" title is deliberate. The company begins by identifying the financial or operational result a client wants and connects its own compensation with achieving that result. Dom argues that many AI initiatives begin without a clear definition of success. The pressure to adopt AI is real, but companies frequently select technology before agreeing on the business problem, desired outcome, or measurement. He identifies three recurring mistakes. The first is framing the project around AI rather than the business need. The second is failing to establish a metric and baseline before work begins. The third is using a consulting model that rewards billable time without connecting payment to the client's result. We also discuss how generative AI is changing traditional build versus buy decisions. Companies historically bought software because custom development was slow, expensive, and difficult to maintain. Coding agents can now reduce the time and cost required to create software for specific internal needs. Dom does not believe SaaS will simply disappear. However, vendors selling convenience, workflow wrappers, or integration glue face new competition from customers who can create similar capabilities themselves. He argues that stronger SaaS positions will depend on assets a model cannot easily regenerate, including proprietary data, networks, regulatory standing, and deep workflow adoption. This leads to a wider discussion about competitive advantage. When companies have access to similar models, generated code begins to converge. Dom believes lasting differentiation comes from company data, institutional knowledge, connected systems, employee experience, and the semantic context surrounding that information. Dom explains why ontologies matter to enterprise AI. Raw data tells an agent what is stored in a particular field. An ontology describes the customers, orders, contracts, payments, relationships, and business rules represented by that data. This context allows people and agents to reason about information in a way that reflects how the company actually works. Governance also needs to be designed from the beginning. Dom argues that security, permissions, accountability, and compliance allow successful pilots to expand without forcing the business to rebuild everything later. How can leaders tell when AI is genuinely being adopted? Dom offers a surprisingly simple signal: people stop talking about AI. The technology becomes part of ordinary Monday morning work, and employees focus on completing the task rather than explaining the tool. Has your company defined the business result, measurement, proprietary context, and governance required to turn AI enthusiasm into operational value? Listen to the episode and share your thoughts with me.

What does a modern contact center need to deliver when the person reaching out may already feel anxious, embarrassed, and unsure where to turn? In this episode of Tech Talks Daily, I speak with Chris Lovell, service delivery lead for StepChange Debt Charity's contact center and product owner for its Genesys Cloud platform. StepChange supports hundreds of thousands of people facing financial hardship each year. Chris explains that approximately 40% of its clients receive Universal Credit, over 60% rent their homes, and many are dealing with an additional vulnerability alongside debt. That context makes the first interaction especially important. People may have delayed asking for support while their financial position became harder to manage. A failed call, long queue, unnecessary transfer, or request to repeat their story can increase stress at the moment they need reassurance and practical help. Before adopting Genesys Cloud, StepChange relied on fragmented contact center technology that experienced regular technical problems and outages. The charity also lacked detailed insight into why clients were making contact at different stages of their journey. People could enter the wrong queue, wait to speak with an advisor, and then discover they needed another team. The technology also affected employees. Chris says frontline colleagues eventually stopped proposing improvements because they did not believe the existing platform could support them. StepChange migrated to Genesys Cloud in four weeks through a three-phase delivery. The team began with lower-risk services, increased the size and complexity during the second phase, and moved the core debt advice operation during the third. Chris says the migration was completed without downtime. The initial objective was to reproduce the existing service on a stable cloud platform before introducing further capabilities. This sequencing gave the team time to correct early issues and adjust training before the largest group of advisors moved across. We discuss how improved intent capture and routing helped one team reduce misrouted calls by 60%. Clients reached the right advisor sooner, avoided repeated explanations, and could move toward a suitable debt solution faster. Advisors also began conversations in the right place instead of apologizing for delays or correcting the journey. Chris argues that contact center success cannot be judged through efficiency alone. StepChange examines whether clients understand their options, complete the advice journey, activate a sustainable plan, and continue toward becoming debt free. The circumstances remain difficult for many clients. Chris says approximately 28% are still in a negative budget after receiving advice, with an average monthly shortfall of around £600. Some conversations require time, empathy, and experienced human support. Around 85% of StepChange advice journeys now happen online. Digital access can offer privacy and flexibility, while advisors remain available for the emotional and complicated moments where a person needs reassurance. We also consider future plans for WhatsApp, web messaging, connected journeys, and AI-powered advisor support. Chris advises leaders to begin with genuine customer behavior rather than selecting a technology and searching for somewhere to use it. How can your contact center remove unnecessary effort while preserving the conversations where people most need to feel heard? Listen to the episode and share your thoughts with me.

What does it really take to build an AI-ready enterprise when your data is fragmented, teams operate in silos, and years of technology decisions have created complexity that no large language model can magically fix? In this episode of Tech Talks Daily, I speak with Raymon Ohmori, Senior Principal Software Engineer at Valiantys, and Jiecheng Dong, Senior Software Engineer at Valiantys, about the work that goes into enterprise AI adoption and why successful AI transformation begins long before companies deploy agents, copilots, or autonomous workflows. Using Valiantys' work with Mercedes as a case study, Raymon and Jiecheng explain how modernizing software delivery and connecting data across teams can create the foundation required for AI systems to deliver meaningful business value. We discuss why fragmented data, organizational silos, poor governance, and unclear business problems continue to prevent many companies from moving beyond AI pilots. The conversation examines what it means to become AI-ready in practice. Jiecheng explains why enterprises need performant, structured, and queryable data rather than simply feeding huge volumes of information into large language models. Raymon shares why businesses must begin with real problems, stakeholder needs, and clearly defined outcomes to justify the cost of AI and successfully move projects into production. We also discuss the growing role of agentic AI and autonomous workflows in software engineering. How should engineering teams prepare AI agents to become active participants in software development? What tools, context, permissions, governance, and observability do these systems need to operate effectively? And why might treating an AI agent more like a new colleague than another software tool help teams think differently about deployment? Raymon and Jiecheng also share their perspectives on AI-assisted software development and developer productivity. As AI becomes increasingly capable of writing code, the role of the software engineer is shifting toward architecture, system design, requirements gathering, trade-off evaluation, and translating business needs into technical specifications. We also discuss the challenge facing junior developers and why companies still need to create pathways for new engineering talent. Finally, we examine the practical steps CIOs, CTOs, and engineering leaders can take today to build more connected, AI-enabled enterprises. From improving data ownership and governance to identifying costly problems that AI can realistically solve, this conversation offers a practical guide for companies trying to move from AI experimentation to production systems that deliver measurable value. Where is your company on its AI journey? Are fragmented data, organizational silos, and unclear business problems preventing your AI projects from reaching production, or have you found effective ways to turn experimentation into measurable results? Share your thoughts with me.   Useful Links   Valiantys Website: https://www.valiantys.com/  Valiantys LinkedIn: https://www.linkedin.com/company/valiantys/    

What happens when AI moves beyond writing emails and summarizing meetings and starts influencing who gets hired, promoted, and paid? In this episode of Tech Talks Daily, I speak with David Lloyd, Chief AI Officer at Dayforce, about why HR is becoming one of the highest-stakes environments for artificial intelligence and how companies can introduce AI while protecting employee data, maintaining human accountability, and preparing for growing regulatory scrutiny. HR systems contain some of the most sensitive information companies hold, from salaries and performance records to benefits and personal data. At the same time, AI is increasingly being introduced across recruitment, workforce management, compensation, performance, and employee experience. David explains why this combination creates enormous opportunities but also places greater responsibility on employers to understand how AI systems operate and how decisions are made. A major theme throughout our conversation is the role of AI governance. David challenges the assumption that governance slows innovation, arguing that the right processes can help companies evaluate AI ideas quickly while reducing the risk of introducing systems that lack appropriate data, transparency, or regulatory safeguards. Dayforce recently achieved ISO/IEC 42001 certification for AI management systems and NIST AI Risk Management Framework attestation. David explains what independent validation means in practice and why companies evaluating AI vendors should ask for evidence of how systems are governed, tested, monitored, and audited. We also discuss the principle of "AI by choice." David argues that CIOs and HR leaders should never discover that a new AI capability has suddenly been activated across hundreds of employees without their knowledge. Companies need visibility into where AI is being used, what data employees can provide to models, and whether customer information is being used to train external AI systems. The conversation examines AI literacy and why HR leaders need to become comfortable with the technology themselves before guiding employees through changes to jobs and working practices. Employees are already experimenting with AI, sometimes through personal tools outside approved company systems. Rather than ignoring this behavior, David explains why companies should provide safe environments where people can learn while establishing clear rules around sensitive data. Human accountability remains central as AI takes on more responsibility. David discusses why people using AI should remain accountable for its outputs and why human oversight matters when technology influences decisions involving recruitment, compensation, performance, and careers. For CEOs, CHROs, CIOs, HR technology leaders, and anyone responsible for enterprise AI, this conversation provides practical guidance on responsible AI adoption, employee data, AI bias, model monitoring, regulatory compliance, vendor selection, and building AI governance that can stand up to scrutiny. The lesson is that governance does not have to be a brake on AI adoption. Done well, it can give companies the structure and confidence to move faster, make better decisions about where AI belongs, and continue using the technology when regulators, employees, customers, and boards start asking harder questions. https://www.dayforce.com/ https://www.linkedin.com/company/dayforce/ https://www.linkedin.com/in/dtlloyd/

What separates an impressive agentic AI demonstration from a deployment that produces measurable business value across an entire company? In this episode, I speak with Frank Theisen, Vice President of IBM Technology across Europe, the Middle East and Africa, about how businesses can move AI agents beyond isolated pilots and into the processes where work actually happens. Frank believes the conversation has changed considerably. Most large companies are deploying some form of AI, yet many still struggle to demonstrate a significant commercial return. The difference comes from connecting AI with end-to-end business processes rather than creating another assistant that sits outside the systems employees use every day. IBM has attempted to prove this internally through its "client zero" approach, using its own technology across human resources, IT, procurement, sales and software development before taking those practices to customers. The company reports that AI, automation and hybrid cloud have contributed to $4.5 billion in productivity gains over three years. Frank explains how IBM's AskHR service handles common employee inquiries and helps managers complete administrative tasks without learning how to operate several separate enterprise applications. IBM reports that AI now resolves 94 percent of common HR requests automatically, while similar work is taking place across IT support and procurement. The discussion then turns to orchestration. As companies acquire agents from multiple software providers, the problem becomes far larger than creating individual assistants. Businesses need to understand how agents communicate, which systems they can access, what identities they use and who remains accountable for their actions. Frank expects the number of applications, agents and non-human identities to grow rapidly. Without orchestration and governance, companies risk recreating the same application sprawl they have spent years attempting to reduce, this time with software capable of making decisions and generating additional code. Data presents another barrier. Publicly trained models rarely contain the proprietary information that gives a company its commercial advantage. That information remains distributed across databases, applications, mainframes and cloud services. Frank argues that enterprises need a governed, federated way to bring AI to their data without repeatedly copying everything into another repository. We also discuss digital sovereignty across Europe and the Middle East. Frank describes sovereignty as a matter of control across data, operations and technology. Companies need to decide which workloads require isolation, which regulations apply and where dependence on one provider could limit their future choices. Wimbledon provides a timely example of these principles in practice. IBM Bob helped modernize the tournament's digital platform by mapping and migrating approximately 15,000 articles, videos, photographs and related metadata. IBM says work that would traditionally require four or five specialists over several months was completed by one engineer within four weeks, with the assets themselves extracted in 47 minutes. Frank closes with three practical priorities. Understand where AI could affect the business, determine how successful use cases can be automated across complete processes, then address security, governance and provider dependence before expanding them. If your company already has dozens of AI pilots, should the next investment create another agent or coordinate the ones you already have? Listen to the episode and share your thoughts with me.

How quickly should an AI investment begin producing measurable business results? In this episode, I speak with Monica Kumar, Executive Vice President and Chief Marketing Officer at Extreme Networks, about the growing pressure on technology leaders to prove that AI investments are producing financial and operational value. The conversation draws on Extreme Networks' State of AI for Networking 2026 report, based on a global survey of 200 C-level executives and vice presidents of IT. The findings suggest that enterprise AI has entered a far less forgiving phase. Experimentation continues, but executives increasingly want evidence that deployments are reducing costs, improving productivity or creating better user experiences. The most striking result is the speed now expected. Some 57 percent of respondents said they expect measurable AI impact within weeks or sooner, compared with 16 percent in the previous year. Projects that once might have received six or 12 months to demonstrate value may now face questions within 30 or 60 days. Monica explains why these demands are changing which AI projects receive attention. Leaders are looking for use cases connected with existing operational problems, where results can be measured and communicated clearly. This makes enterprise networking an interesting test case. AI workloads depend on network compute, bandwidth, availability and access to current data. According to the research, 92 percent of respondents said AI is increasing demands on network compute and bandwidth. A fragmented or outdated network may therefore limit the performance of the AI applications running across it. The network can also provide an early opportunity to show what AI produces in practice. Monica discusses performance monitoring, predictive analysis, troubleshooting, compliance checks, capacity planning and security. These are repetitive, data-heavy activities where improvements can be measured in time saved, fewer support tickets and better service availability. A case from Middlesbrough College brings those claims into focus. The college reports that firmware tracking fell from as much as five hours each week to approximately five minutes, while the time spent troubleshooting decreased by around 90 percent. For a small network team, the value comes from giving people additional capacity without asking them to monitor every device or event manually. We also discuss why AI capabilities work better when embedded within normal business systems rather than added as another standalone tool. If an AI service remains outside the daily workflow, employees must move between platforms, transfer information and interpret the result themselves. Integrated AI can monitor the network, identify anomalies, recommend action and automate routine work within the environment where the team already operates. Monica also warns that the quality of AI depends heavily on its data. Before investing in another model or application, businesses need accurate, current and well-managed information. They must also examine whether their network has the capacity to support additional workloads and whether employees understand how to use AI responsibly. If executives expect AI results within weeks, are businesses selecting the right use cases, or simply imposing unrealistic deadlines on complicated technology programs? Listen to the episode and share your thoughts with me.

Could the disaster recovery plan designed to protect your company make a ransomware incident even worse? In this episode, I speak with Darren Thomson, Vice President and Chief Technology Officer for EMEA at Commvault, about Resilience Operations, commonly known as ResOps, and why cyber recovery now requires security, infrastructure, identity and data teams to work from one coordinated plan. Darren argues that many companies are accepting a difficult reality. Even with considerable investment in prevention and detection, a breach may eventually succeed. That does not make cybersecurity controls any less necessary, but it means recovery can no longer be treated as a secondary activity managed by another department. The problem is that security operations and infrastructure teams have traditionally worked toward different objectives. Security specialists concentrate on identifying and stopping threats. Infrastructure teams protect data, maintain backups and restore systems after outages. During a cyberattack, a successful recovery requires both sets of expertise. A backup administrator may be able to restore data quickly, but a forensic specialist must establish whether that data is clean. Without that confirmation, the company risks restoring malware and restarting the incident. Darren explains why a conventional disaster recovery plan may be particularly dangerous during ransomware. These plans were commonly designed for physical failures such as a lost data center. Data would be copied from one location to another so operations could continue. If the source data is infected, however, fast replication can carry the malware into the recovery environment. This is where ResOps enters the discussion. Darren describes it as an operating model rather than a product. It combines established practices from security and infrastructure management into a continuous program for testing, learning and improving recovery. Individual technology projects may come from the program, but resilience itself never reaches a final completion date. AI adds pressure on both sides. Criminals can use it to create faster and more effective attacks, while defenders can use machine learning to inspect large volumes of information, detect patterns and identify the newest clean recovery point. Companies must also protect AI systems as they would any other business application, including the models, data repositories and identities connected with them. Darren offers one practical starting point for CIOs and CISOs: Mean Time to Clean Recovery, or MTCR. This measures how long it takes to restore an application and its data with evidence that both are free from compromise. Before measuring MTCR, leaders must define their minimum viable company. These are the systems and services the business cannot operate without. Once that list exists, teams can test how long a verified clean recovery would take and replace assumptions with evidence. The initial answer may be uncomfortable. Teams may know how to restore an application without knowing whether the backup is clean. Security may know how to inspect the system but lack an established workflow with the recovery team. Darren sees those gaps as the starting point for a useful ResOps program because they provide everyone with a shared problem and a measurable objective. If your most important systems disappeared today, how long would it take to bring the minimum viable company back using verified clean data? Listen to the episode and share your answer with me.