Your team uses AI every day. Is it paying off?
Logins tell you AI is switched on. They don’t tell you if the work got better. I find where AI pays off in your team, build it into the work, and show you what changed in numbers your leadership will trust.
Two-time founder (AuthMetrik, Tatsam) with 12 years in product. Machine learning since 2017, from fraud detection backed by a YC Startup School grant to an AI health assistant at Medanta. Now I lead AI adoption end to end, from choosing the work to measuring what changed.
How I tell if AI is working: the Token Line
Most AI tools will happily tell you how many people logged in. Nice to know, but it doesn’t tell you whether anything got better. So I put together a simple way to look past the login numbers. I call it the Token Line.
How we can work together
AI value assessment
A close look at how your team uses AI today, where it's helping and where it isn't. For each use case, you'll know whether to expand it, improve it or stop it.
Claude, ChatGPT or Copilot rollout
You've paid for the licences. Let's make them earn their keep. I set the tools up around how your team works, and we check the results at 30 and 90 days.
AI adoption programme
We pick the right tasks, build the tools, and I work with your people until the new way of working sticks. We finish with a before-and-after look at the work.
Fractional Head of AI
If you need someone senior to own AI but aren't ready to hire full-time, I can be that person until you are.
What I'm curious about right now
I want to know whether AI delivers what teams bring it in for, and what it quietly costs in review time and skills. I write about what I find in Build and Beyond, my newsletter, every two weeks. Each issue comes with a tool you can try the same day.
Every two weeks. No spam, vendor endorsements, or breathless hype. Unsubscribe any time.
Good mentors make better AI users
Some managers expect AI to read their minds. They've been expecting the same from their teams for years.
6 OCT 2026If not tokens, then what?
A six-layer test for whether AI is paying off.
30 SEP 2026AI does half the job. Who's doing the other half?
Your team's AI work looks finished. Here's why you can't tell if it is, and a quick diagnostic to find the missing half.
Why I care about AI safety
Healthcare taught me what's at stake. Therapy done badly can leave someone worse off than no therapy at all, and no technology replaces a person you trust.
AI feels like the same story at a far larger scale. Used well, it takes the busywork off our hands and gives us time back for the people we love.
Used carelessly, it wears down something much harder to rebuild: our own judgement. People already take its answers on faith. Some lean on it as an emotional companion, and reports of AI-induced psychosis are starting to surface. As models grow more capable, the cost of getting this wrong grows with them.
We're still early. Most people are learning to use AI right now, and the habits they form this year will be hard to undo. That's the window I want to work in.
"I'm looking for ways to go deeper, through residencies, fellowships and work with researchers who study how people trust and oversee AI. If that's you, I'd love to talk."
Used carelessly, AI wears down something much harder to rebuild: our own judgement.
Is your AI paying off? Let’s find out.
Tell me what your team brought AI in for and where it’s stuck. In 30 minutes we’ll work out whether it’s paying off and what a sensible first piece of work looks like. You’ll get a written proposal within 48 hours.