Tech Talks Daily

What would make you comfortable giving an AI agent the authority to act inside a live business process?
In this episode of Tech Talks Daily, I speak with Valerie Henderson, CEO of Caylent, about what has to happen between a successful demonstration and a system a business can depend on. The conversation centers on production readiness, including access controls, spending limits, traceability, workflow design, and the people responsible for the outcome.
Valerie describes why introducing agents requires changes to how work gets done. If an organization adds new technology without reconsidering the handoffs and decisions around it, the hoped-for productivity gains can remain difficult to achieve. Her advice is to begin with a bounded problem where automating a few steps could provide a clear benefit, then build on the evidence rather than starting with an ambitious network of agents.
We discuss the difference between read-only workloads and agents that can alter live systems. Valerie says many customers are already comfortable with agents retrieving information, while autonomous operations such as self-healing infrastructure are concentrated among the most advanced customers she sees. Those observations reflect her experience with Caylent engagements. They do not establish a universal level of enterprise adoption.
For me, one of the most useful questions is what meaningful human oversight looks like. Valerie is clear that people remain accountable for the instructions and access they give agents. She emphasizes deliberate testing, appropriate governance, and the ability to understand what a system has done. We also discuss how agents can repeat work without producing anything useful, creating a bill that needs to be considered alongside accuracy and reliability.
The conversation moves into Caylent's own operating model. Valerie describes moving away from assumptions built around human time and materials, making existing technical knowledge easier to reuse, and changing delivery incentives to reward speed and customer satisfaction. It is a practical reminder that organizations adopting AI may need to reconsider what they reward as well as which tools they buy.
She also shares a customer example involving Smarsh. According to Valerie, redesigning its software delivery process with agentic AI freed engineering capacity that could be reinvested in the business. She describes that capacity in financial terms, but the interview does not establish a measurement period or separate capacity value from realized cash savings. The useful lesson is the connection between a defined objective and a decision about how to use the resulting capacity.
We touch on Caylent's agentic AI readiness research, while keeping the broader discussion focused on practical experience. The useful question for leaders is what controls and evidence should come before greater autonomy, and how they will reassess that decision as the system changes.
Finally, Valerie explains how she thinks about success for employees as well as customers. She wants technologists to have time for difficult problems and opportunities to develop their skills. She also argues that customers should be able to explain the value of the work they have undertaken. Model performance is part of that assessment, alongside the environment, permissions, costs, and business process in which the agent operates.
I would love to hear how your team is making these decisions, so please share your thoughts. Which responsibilities are you ready to delegate to an AI agent, and what evidence would you need before going further?

Useful Links
Caylent website: https://caylent.com/

Caylent's Agentic AI Readiness survey report: https://caylent.com/blog/agentic-engineering-survey-report

What makes an AI fall alert useful to the person who needs help and the team responsible for responding?
In this episode of Tech Talks Daily, I speak with Mike Link, chief operating officer at KamiVision, about deploying computer vision in residential care. We discuss the connection between the sensor, the people receiving its notifications, and the decisions that follow. The conversation also looks at privacy, connectivity, false alarms, and why a system that works in a care facility needs further work before it can serve a wide range of homes.
Mike describes KamiVision's path from home-security cameras to its KamiCare platform. His account of fall detection involves three stages. A device identifies a possible event, a cloud model filters notifications, and a person reviews a blurred, encrypted clip before an alert reaches the user. The human review is part of the service he describes, so it belongs in any discussion of how the system performs.
I ask about the company's 99.5% accuracy claim and what buyers should understand before relying on a number like that. Mike explains the filtering process, but the interview does not establish a test population, a missed-event rate, or a complete statistical definition. His answer is a starting point for further questions about performance across the whole service, rather than proof that a device detects 99.5% of all falls by itself.
One of the most useful examples concerns a resident found on the floor. Mike describes how reviewing an event can help staff understand whether someone fell or deliberately lowered themselves because they needed help with a task. That information may lead to a change in a routine or room layout. It is an example of how visual context can inform care, rather than an independently measured claim that the technology prevents injuries.
We also discuss the difference between a managed care setting and a home. A residential facility can have relatively consistent room layouts and lighting. Homes introduce furniture, pets, several rooms, and areas outside a sensor's view. Mike talks about investigating other sensing technologies to address those gaps. He presents those ideas as work under consideration, not a set of generally available capabilities.
Connectivity is another practical issue. Mike says the version discussed uses Wi-Fi and that a cellular option has been considered. He also explains that the consumer version was being tested with families and was not yet available to purchase at the time of recording. That distinction between an established care-facility offering and a product still being developed matters when assessing where the technology can be used today.
Privacy and consent are part of deployment. Mike says facilities obtain consent from residents or an appropriate representative, and he describes the service as focused on event clips rather than live viewing into a room. Those explanations help frame the conversation a buyer should have about access, retention, review, and the information people are agreeing to share.
Finally, we discuss the business case. Mike connects the value of event information with staff response, communication with families, and adjustments to care. He also describes possible reporting and risk-identification capabilities that had not been deployed. The interview offers a view of the operational questions involved, without establishing a controlled clinical result or a fully quantified return on investment.
I would love to hear your thoughts, so please share your experiences. How should care providers assess whether an AI alert improves the work people do, while respecting the privacy and choices of those receiving care?

What can a crypto exchange's proof of reserves establish, and what might still be missing from the picture?
On Tech Talks Daily, I'm joined by Harvey Liu, co-founder and CEO of LeveX, to discuss transparency, platform engineering, and the responsibilities that come with operating a centralized exchange. Harvey brings a computer science background, an MBA, and experience in venture capital. Our conversation looks at how becoming a founder changed the questions he asks about growth, spending, product delivery, and customer assets.
We begin with the community idea behind LeveX. Harvey recalls meeting other crypto investors and wanting an online environment where people could discuss ideas alongside trading. He describes that as part of the motivation for building the company. The interview also makes clear that some social features remain in development or are planned for a later release. A community vision and a finished product roadmap are different things, and the discussion gives listeners a view of that work in progress.
The central technology question is proof of reserves. Harvey explains why a snapshot should be considered in context, including its date, the assets it contains, and the liabilities it may not reveal. He discusses the limitations of relying on a platform's own token and the difficulty of understanding obligations outside the information shown. Publishing reserve information can contribute evidence, but the conversation does not establish that a snapshot alone proves a business is solvent or that customer assets are protected under every condition.
We also discuss the engineering approach Harvey says LeveX uses. He describes Merkle-tree verification, user account inclusion, and separating parts of the reporting process from the live trading environment. The challenge is to provide useful information while managing privacy and system load. His explanation is a founder's account of the architecture, rather than an independent technical audit or a live demonstration.
From there, the conversation turns to the realities of building the business. Harvey says he underestimated how much product development investment would be needed to compete with larger teams. He also describes the difference between taking a minority investment position and being responsible for the operation itself. Cash flow, customer growth, infrastructure decisions, and the consequences of a failure become immediate concerns rather than topics for a periodic review.
Security and AI bring further tradeoffs. Harvey describes security reviews, operational controls, and the ability to suspend certain functions if something appears wrong. He also discusses AI-assisted development and analysis, alongside agents being built to help with order placement. These ideas raise questions about access, authorization, accuracy, and responsibility. The interview does not demonstrate those systems or establish that AI-generated analysis improves trading results.
Availability is another part of the discussion. Harvey says LeveX restricts access from jurisdictions including the United States, the United Kingdom, Singapore, and mainland China. He describes licensing work as being in progress, rather than announcing completed authorizations. This episode examines a technology business and its founder's decisions. It does not recommend opening an account or imply the service is available in every listener's location.
We close with Harvey's views on decentralized exchanges and plans for additional social features, competitions, and copy trading. These remain company plans as described at recording, without a verified delivery schedule or a promise of financial returns. For listeners interested in platform design and founder accountability, the useful thread is how to question evidence, distinguish current operations from future ambitions, and assess what a company is actually claiming.
Please share your thoughts. What should exchanges make easier to understand about their reserves, liabilities, and operating controls so people can assess their claims with greater clarity?

What needs to happen before an IT team can trust an AI agent to resolve a problem rather than simply report it?
In this episode of Tech Talks Daily, I welcome Jed Ayres, CEO of ControlUp, and Gadi Feldman, Chief Product Officer, to discuss autonomous endpoint management and the practical work behind it. We begin with a familiar problem for enterprise IT teams, plenty of monitoring data and recurring issues that still require investigation, ticket transfers and manual intervention. Jed explains why the aim is to follow an operational problem through to an authorized action and then verify that the issue has been resolved.
Gadi describes how specialist agents could bring together expertise across areas such as endpoint management, networking and storage. We discuss the connections those agents need to existing tools and why an understanding of the technology is only part of the job. Agents also need access to company procedures, priorities and knowledge, much as a new member of an IT team needs to learn how the organization works. Gadi explains how configurable human approval and other controls form part of that approach, with teams able to begin with analysis and recommendations before allowing additional actions.
Jed shares a CHRISTUS Health example involving an audio driver issue across approximately 900 devices, illustrating the value of seeing a common pattern rather than treating every device as a separate support problem. We also discuss what a useful business case should measure. Alongside the time spent by IT, leaders need to consider the employee's experience of waiting for help or finding a workaround. Device performance and software usage can also inform decisions about equipment replacement and unused licenses.
The conversation considers what becomes possible when repetitive work takes less of the team's time, while acknowledging the need to establish trust and define the limits of autonomy. For leaders looking for a starting point, Jed describes ControlUp's seven-day challenge, and Gadi recommends understanding the fleet already in place. Which recurring IT problem would you address first, and what evidence would convince your team that automation was helping? Please share your thoughts with me.

When would you trust your ERP system to move from answering a question to carrying out the work?
Recording at SuiteWorld in Las Vegas, I spoke with Brian Chess, who leads AI product and technology for Oracle NetSuite, about what that progression looks like in practice. Brian was an early developer when the business was called NetLedger, later left to build a company, and returned with a different perspective on the needs of small and growing businesses. Our conversation focuses on how AI can help people understand their systems, complete unfamiliar tasks, and gradually delegate work without losing control of the process.
One of Brian's most useful ideas is AI acting as a teacher. He describes users asking Ask Oracle who has logged into their account, or asking it to explain a script they do not fully understand. The benefit is the answer and the explanation of how to find it. Someone who previously needed help from a colleague can begin to understand a part of the system for themselves.
He also shares an example involving NetSuite customer MCNC. When the payroll administrator was unavailable, Brian says a controller who had not been trained on the process used Ask Oracle for guidance. The outcome he describes is straightforward and meaningful: payroll ran on time. It is a customer account from the interview, rather than an independently assessed result, but it illustrates how useful AI assistance can be without requiring a dramatic productivity claim.
Financial close provides another practical example. Brian describes starting with the spreadsheet a finance team already uses, modelling that process inside NetSuite, and introducing automation in stages. I recognise the appeal of that approach from my own time supporting month-end and year-end work in IT. Teams need to understand how the process will run, especially when a mistake can affect reporting or the system of record.
We discuss how permissions apply as software gains the ability to act. Brian says Ask Oracle operates with the permissions of the user, while an agent receives an assigned role governing its access. He describes a progression from read-only analysis to limited changes as users gain confidence. Human approval remains available for transactions, particularly while teams refine instructions and account for cases they may have missed.
Business context and shared knowledge are part of the same discussion. Brian explains how written skills can capture procedures and the intended business outcome, making knowledge available beyond the colleague who usually handles the task. We also talk about accountability. He describes transaction records showing which agent acted and when, together with an activity log. Those records help establish what happened, although they should not be confused with a complete explanation of every internal model decision.
The conversation also covers whether to build AI capabilities or use those embedded in a business platform. Brian's answer depends on the role the capability plays in a company's competitive position, the expertise available, and the time it wants to invest. His view of ROI is similarly specific to the business. Comparing vendor contracts with billing records may reveal savings, but the same result does not necessarily transfer to a company with different suppliers and arrangements.
We finish with the opportunity Brian sees for smaller businesses that have fewer specialists available. AI assistance may help people undertake work they previously needed a consultant or experienced colleague to explain. The question is how to extend that access while retaining appropriate permissions, review, and responsibility.
I would love to hear your thoughts, so please share your experiences. Where would conversational ERP make the biggest difference in your business, and which actions would you still want a person to approve?