
Independent AI due diligence and advisory for investors, boards and acquirers.
LondonHybrid
I run Agathon, an independent AI advisory practice in London. My work is judgement for expensive, hard-to-reverse decisions: the read you get before you spend the money.
Most of my time goes on technical due diligence for acquirers, investors and boards looking at an AI company or an AI claim. The question is whether the AI under the valuation is real. Answering it means reading the model, not just the code around it: reproducing the headline result, testing the benchmark the deal rests on, and looking for the failures a code review does not reach. I lead every engagement myself, work to a fixed fee, and go in with a mandate to falsify rather than confirm. I have no build arm and nothing to sell afterwards, so the finding is the only product.
The rest of my work is advisory. I work with leadership teams making AI commitments on which claims to trust, what to buy, what to build, and how to know whether it worked. That includes fractional CTO work for organisations taking an AI product to market, and preparing boards and executives who have to defend an AI decision.
Before Agathon: a PhD in computational linguistics from Cambridge, research in natural language processing, and a CTO tenure taking an AI product to public beta.
Members write their own listings. An organiser checks that they are a real member; that is not a quality mark, and we do not verify what a listing claims. Tell us if something here is wrong.