What happens when AI can browse, understand, and interact with the web just like a human?

This week on the ThinkData Podcast, I'm joined by Abhishek Das, Co-founder and Co-CEO of Yutori.

Before founding Yutori, Abhishek was a research scientist at Meta, where he worked on AI agents long before the recent explosion in generative AI.

Today, Yutori is building specialised AI models designed specifically for web agents—systems capable of navigating websites, completing tasks and operating autonomously in complex online environments.

In this episode we discuss:

• Why specialised AI models outperform general-purpose LLMs for web automation

• The biggest technical challenges of building reliable AI agents

• Why AI demonstrations often fail in production

• Product-market fit in one of AI's fastest-moving markets

• What founders consistently underestimate when building AI companies

• How AI agents could fundamentally change the way businesses and consumers use the internet

If you're interested in AI infrastructure, startups, product, engineering, or the future of autonomous software, this is an episode you won't want to miss.

Enterprise data has never been more valuable, yet most companies still struggle to answer simple business questions.

This week, I sat down with Ethan Ding, Co-Founder and CEO of TextQL, to discuss why traditional analytics tools weren't built for the age of AI, how autonomous agents are changing enterprise infrastructure, and what happens when AI starts generating thousands of queries where humans once generated dozens.

We discuss:

• Why today's BI tools are reaching their limits

• The biggest technical challenges behind enterprise AI

• How TextQL is rethinking the modern data warehouse

• Why data analysts aren't disappearing—but their jobs are changing

• Finding product-market fit in one of AI's most competitive markets

• What enterprise software looks like over the next five years

If you're building, investing in, or buying AI products, this is a conversation you won't want to miss.

Agentic engineering is quickly becoming one of the most important shifts in software development.

In this episode, I sit down with Scott Breitenother, Co-Founder & CEO of Kilo Code, to discuss why individual coding assistants won't drive the future of software development, but by autonomous AI agents capable of planning, building, testing, and shipping software.

Scott shares the journey from building and selling Brooklyn Data Company to launching Kilo, an open-source agentic engineering platform designed to help developers become dramatically more productive in the AI era.

We discussed product-market fit, engineering adoption, the realities of competing with Cursor, Copilot, and Claude Code, and what engineering leaders should be thinking about as AI fundamentally changes how software teams operate.

If you're a founder, engineering leader, developer, or simply interested in the future of AI-powered software development, this is an episode you won't want to miss.

In this episode of the ThinkData Podcast, I welcome back Dimer Health Co-Founders Carrie Hodge, Sarig Reichert, alongside Chief Medical Officer Dr. David Feldman.

Fresh off the announcement of their $13.5 million Series A, the team discusses why the 30 days following hospital discharge remain one of the most broken parts of healthcare, and how Dimer is using clinician-led AI to transform patient outcomes.

The conversation covers scaling after fundraising, building culture during hypergrowth, the realities of operating at Series A, and what it takes to create a new layer of healthcare infrastructure.

Today, we’re getting into where AI actually fits in the finance stack and why most teams are getting it wrong.

I’m joined by Deepak Bapat, CTO and co-founder of Tabs, an AI-native platform automating contract-to-cash end-to-end.

We cover:

• The difference between AI-native vs AI-layered products
• Why finance has near-zero tolerance for AI errors
• The technical challenge of automating contract-to-cash workflows
• Integrating AI into fragmented ERP ecosystems
• What finance teams may look like in the next 3–5 years as AI agents take over operational workflows

A really interesting conversation around where AI creates real operational leverage, and where the hype still outweighs the reality.

Most enterprise AI isn’t failing because of the models…

It’s failing because no one trusts the output.

In this episode of the ThinkData Podcast, we’re joined by Uri Knorovich, CEO of Nimble, to unpack the growing trust gap in enterprise AI and why bad, delayed, and fragmented data is becoming the biggest blocker to real AI adoption.

We explore:

• Why AI works in demos but breaks inside enterprises

• The hidden risk of AI outputs without traceable data lineage

• Why trusted, live data infrastructure matters more than model speed

• The shift from “is the AI fast?” to “can we trust the decisions?”

• Real-world examples where AI and live data are already driving value

A great conversation on the future of enterprise AI, data infrastructure, and building systems companies can actually rely on.

On today’s episode, we’re joined by Sadegh Salehi, Director of Research and Principal Scientist at Overjet, a Series C AI-powered dental platform spun out of MIT.

Overjet is using computer vision and AI to analyse dental X-rays, helping dentists, insurers, and patients make more accurate, transparent, and data-driven care decisions.

In this episode, we explore:

• The hardest machine learning challenges in dental imaging

• Building AI systems clinicians and insurers can trust

• Why Overjet expanded from software into hardware

• What it really takes to drive AI adoption inside clinical workflows

• The future of AI in dentistry over the next 5–10 years

A fascinating conversation on applied AI, healthcare infrastructure, trust, and the future of medical imaging.

Most AI today is intelligent… but it doesn’t feel human.

In this episode, I sit down with Vishnu Hari, founder of Ego AI and former AI researcher at Meta, to unpack what’s missing and why “humaneness” might be the biggest unlock in AI.

We cover:

• Why AI still feels robotic
• The gap between intelligence and behaviour
• Why most AI companies are solving the wrong problem
• And where the real opportunity sits

If you think AI is already “good enough”, this will challenge that.

Today’s guest is Anjan Kundavaram, Chief Product Officer at Fivetran, a company at the heart of the modern data stack.

In this episode, we unpack one of the biggest questions in AI right now

Who actually controls customer data

We explore what data ownership really means in the age of AI, how API restrictions are shaping what companies can and cannot build, and what happens to innovation if data becomes harder to access

We also dig into how Fivetran thinks about “data freedom” and what enterprises should be doing today to stay ahead

If you are building, investing, or operating in AI, this is one you do not want to miss

In this episode of ThinkData, I sit down with Wiley Jones, CEO and Co-Founder of Doss.

Doss is a Series B, hypergrowth startup rethinking ERP from the ground up. They are building a modern, AI-native operations stack for physical product companies, designed to replace spreadsheets, legacy systems, and slow, painful implementations.

We cover:

• Where legacy ERP vendors like NetSuite and SAP have fundamentally missed the mark
• Why now is the moment for a new category of operations software
• What makes Doss a credible replacement, not just another layer
• How AI is actually being applied inside ERP today
• Why traditional ERP implementations fail and how Doss is changing that
• The hardest product and GTM challenges they are solving right now

If you are building or scaling an operations-heavy business, this is a must-listen.