Tech Talks Daily

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.