Mark Rochefort

This is the most exciting time I've had since the World Wide Web kicked off in the 90s. Back then I was doing an MSc in Computer Science at Bristol - writing papers on autonomous software agents (one was called ‘“It's life Jim but not as we know it.” Artificial Life: The Final Frontier in the Search for Artificial Intelligence?’), building an online record shop, and wondering what the web would become. It was 1998. That feeling is back.

I've spent 27 years building digital things: CD-ROMs and the BBC's first push to get people online (Webwise), interactive TV (the Big Brother stuff I worked on won a BAFTA and a few other awards), then eight years heading up technology at Harvest Digital - where I learned that the interesting problems were always about marketing, data, and how people actually make decisions. I co-founded Measurelab in 2013 and spent twelve years in marketing analytics on the Google stack - tracking, tagging, data products, cloud infrastructure, engineering, analysis and more.

Now AI is transforming that space beyond recognition. It's exciting. It's scary. I'm leaning into it.

In 2023 we started building Brian, an all-company conversational AI agent at Measurelab. We learned what works, what hallucinates, and what governance you actually need before you trust an agent with real data. In January 2026 I took the essence of that and made it personal - gave an AI agent real freedom, not as a product, but as a way of becoming unreasonably effective myself. It stuck around. Now it runs my mornings, processes my meetings, holds my thinking, and maintains its own website.

Everyone's struggling with the same thing: too much data, not enough trust in it, and no clear way to make AI work reliably on top of it. The scale is different. The problem isn't.

I believe the future of expertise isn't individual. It's collaborative. Not human or AI. Something in between that neither of us could do alone. Governance from above. Autonomy from below. That's where the leverage is.