The Tech Trek

Raj Koo, CTO at DTEX, joins The Tech Trek for a sharp conversation on insider risk, shadow AI, and why security teams need a more modern way to think about intent. This episode is worth your time if you are trying to understand how AI is changing cyber risk, why non malicious behavior can still create major exposure, and what it takes to protect the business without slowing down innovation.
Raj explains why the old approach of blocking known bad behavior is no longer enough. As employees bring personal AI tools into the workplace, security teams are dealing with a new reality, one where productivity gains, agentic workflows, and data exposure are all colliding at once.
In this episode
Why DTEX focuses on inferring intent, not just catching exfiltration

Why shadow AI is different from shadow IT, and harder to control

How non malicious employee behavior can become the biggest insider risk category

Why agentic AI raises the stakes for visibility and governance

How mature insider risk programs are shrinking response times even as costs rise

Timestamped highlights
00:00 Raj Koo on inferring intent in cybersecurity

01:59 Why early warning signals matter more than the exfiltration point

04:38 The rising cost of insider risk

06:25 How shadow AI became a major non malicious risk

08:13 Why shadow AI is more complex than shadow IT

17:53 Detection times are improving, but the cost problem is getting worse

Standout line
Security has a chance to stop being seen as the function that blocks productivity and start being seen as the function that helps the business adopt better tools safely.
Practical takeaway
If your team is dealing with AI adoption in the wild, start with visibility before judgment. Understand which tools people are using, what they are using them for, and where the real risk sits before defaulting to blanket restrictions.
Link to 2026 Cost of Insider Risks Global Report: https://ponemon.dtex.ai/
Follow The Tech Trek for more conversations with builders, operators, and technology leaders shaping how modern companies work.

Sumeet Arora, Chief Product Officer at Teradata, joins The Tech Trek for a sharp conversation on the shift from human driven SaaS to agentic software. This episode digs into what changes when software stops just supporting human workflows and starts driving outcomes alongside people, why trust and governance matter more as AI systems take on more responsibility, and what serious companies need to do now to prepare.
This is a practical discussion about where the market actually is, what gets overhyped, and what leaders should focus on beneath the noise. Sumeet lays out a clear view of the emerging enterprise stack, from knowledge and context to agents, governance, and outcomes. He also explains why the winners may not be the loudest companies in AI, but the ones that get their data, knowledge, and operating model right.
In this episode
• Why agentic software is a real shift, but still in its early stages• What trust, governance, and explainability need to look like in an AI first enterprise• How software companies should rethink product strategy for agents as well as humans• Why every employee may need to become a manager of AI agents• Why knowledge infrastructure could matter more than the agent layer itself
Timestamped highlights
• 00:45 Teradata’s role in helping enterprises become autonomous• 02:34 Where we really are in the agentic AI maturity curve• 10:16 How software shifts from workflow centric to outcome centric• 16:17 Why every employee may need an AI workforce• 21:57 The skill gap between enterprise users and agentic adoption• 24:48 Why knowledge, not just agents, will define the winners
Standout line
“The fundamental winners will be ones who get the knowledge fabric correct.”
Practical takeaway
If you are building for an AI driven future, do not start with agents alone. Start with trusted knowledge, usable context, clear policies, and systems that can explain decisions. The companies that treat agentic AI as a stack, not a feature, will be in a much stronger position.
Follow The Tech Trek for more conversations with leaders shaping the future of technology, product, AI, and enterprise transformation.

Victor Fang, CEO and Founder of Anchain AI, joins The Tech Trek for a timely conversation on crypto crime, AI driven fraud, and what financial institutions need to understand as digital assets move closer to the mainstream. This episode is worth your time if you care about cybersecurity, compliance, crypto risk, anti money laundering, or where agentic AI is starting to reshape investigation work.
This conversation goes beyond headlines. Victor breaks down how bad actors are using generative AI for phishing, identity fraud, exploit development, and ransomware, then explains how defenders are using AI, graph intelligence, and agent workflows to fight back. It is a sharp look at the collision of crypto, cybersecurity, regulation, and AI infrastructure.
In this episode
What crypto crime actually looks like today, from exchange hacks to romance scams and ransomware

Why crypto risk now extends well beyond crypto native users

How financial institutions, regulators, and compliance teams are adapting

Where AI is helping attackers move faster, and where it is giving defenders an edge

Why agentic workflows and MCP powered investigation tools could change this category fast

Timestamped highlights
00:00 Victor Fang on crypto crime, AI versus AI, and agentic AML

00:53 What Anchain AI does and why blockchain investigation is becoming more important

01:56 How generative AI is already being used in crypto crime and phishing

06:30 What banks, regulators, and AML teams need to understand about crypto adoption

10:44 Why Victor believes AI can give defenders the advantage

16:17 How Anchain uses blockchain data, graph intelligence, and agent workflows to investigate faster

22:04 Why the company’s MCP server could extend beyond crypto into KYC and financial applications

25:21 What the next wave of agent driven security and investigation might look like

One standout idea from the conversation, crypto is much closer to you than you think.
Practical takeaways
Crypto risk is no longer a niche issue, it is increasingly tied to broader fraud, ransomware, and financial crime

AI is accelerating both offense and defense, which raises the bar for security and compliance teams

Agentic investigation workflows could dramatically reduce manual work in AML, fraud, and cyber operations

Companies building in regulated spaces need infrastructure that can handle both speed and scrutiny

Follow The Tech Trek for more conversations with builders, operators, and technical leaders shaping what comes next.

Cam Crow, Director of Data and Analytics at Vacatia, joins The Tech Trek to unpack what happens when a startup outgrows informal ways of working. This episode looks at how data teams can introduce project management frameworks without killing speed, how to manage stakeholder demand as complexity rises, and why the right operating model matters even more as AI begins to reshape analytics work.

Cam shares a practical view from the middle of real growth, from startup scrappiness to acquisitions, migrations, and a much wider stakeholder base. He explains when process becomes necessary, how to build trust during that shift, and where AI is starting to change both delivery workflows and the future of business insights.

In this episode
• Why early stage teams should add process cautiously, not by default
• The moment speed and quality start breaking under too many competing requests
• How public communication and domain based stakeholder channels reduce friction
• Why planning routines matter as much for stakeholders as they do for the data team
• Where AI fits today, from faster delivery to semantic layers that support better answers

Highlights
00:00 Cam Crowe joins the show to discuss project management frameworks through the lens of data, startup growth, and stakeholder alignment
01:58 Why Cam resisted formal sprint planning in the startup phase and why that made sense at the time
05:58 The tipping point where too many priorities start hurting both velocity and quality
11:49 How moving conversations out of direct messages and into domain channels changed team operations
15:03 Inside the two week development cycle and the planning week that keeps stakeholders engaged
21:08 How Cam is thinking about AI, semantic layers, and the future of on demand analytics

A standout idea from this conversation, process should be added conservatively, only when the business truly needs it.

Practical takeaways
• Do not formalize too early, but do not wait until the system is already breaking
• Make prioritization visible once demand exceeds capacity
• Use shared channels instead of one to one communication to reduce bottlenecks
• Build stakeholder rituals into the operating model, not just team rituals
• Treat AI readiness as an infrastructure challenge, not just a tooling decision

Follow The Tech Trek for more conversations with operators, builders, and technology leaders shaping how modern teams work and scale.

Deep Sogani, SVP and Group Data Management Officer at Datasite, joins The Tech Trek to unpack why data governance, lineage, and business process design have become mission critical in the age of AI. This conversation gets past the surface level AI hype and into the operational reality, how companies actually build trustworthy systems, where AI initiatives break down, and why strong data foundations now shape business outcomes in real time.
This episode explores the shift from downstream analytics to data that actively drives live decisions, workflows, and automation. Deep explains why many AI projects fail before the model even matters, how business architecture should lead technical design, and why human oversight still matters in high stakes environments.
In this episode
Why AI has made data governance and data lineage far more operational

Why business process clarity matters before data architecture or tooling decisions

How real time AI changes the demands on data quality and system design

Where agentic AI fits, from workflow automation to more advanced decision support

Why human judgment still matters in AI systems shaped by risk, ethics, and security

Timestamped highlights
01:47 Why AI raises the stakes for governance, lineage, and trust in data

04:57 Why business architecture has to lead before technical design

09:11 The progression from predictive models to agentic AI workflows

17:55 Why the human in the loop is still essential

21:16 What makes an AI project worth prioritizing

26:06 What has changed, and what has not, in AI related change management

Standout line
“Business architecture and business thinking should dictate the what and the why, and the data architecture is the how part which needs to follow.”
Practical takeaway
If you are evaluating AI inside the enterprise, do not start with the tool. Start with the business problem, the workflow, the decision risk, and the quality of the data behind it. Strong models on the wrong problem still fail.
Follow The Tech Trek for more conversations with leaders shaping technology, data, AI, and the future of modern business.