Skip to content

Gregg Bayes-Brown

Founder, Bayes Ltd

Ex-journalist. AuDHD. Bayesian thinker. I started Bayes Ltd because most AI initiatives fail for a reason nobody wants to talk about.

I spent years as a journalist — the kind where you sit across from someone, find the thing they're actually saying underneath the thing they think they're saying, and write it down before they change their mind. Turns out that's useful training for working with AI. Both require the same skill: knowing what question to ask and being honest about what the answer means.

I'm AuDHD, which is a fancy way of saying my brain pattern-matches across domains whether I want it to or not, and occasionally forgets where I put my keys whilst solving a problem nobody asked me to solve. The autism gives me systems thinking. The ADHD gives me the lateral connections. Together, they make me annoyingly good at seeing how things fit together — and completely useless at small talk.

I founded Bayes Ltd in February 2026 because I kept watching the same thing happen. Companies would buy AI tools, run a training day, and then wonder why nothing changed. The answer was always the same: they'd upgraded the software without upgrading the thinking. A better tool in the hands of someone who doesn't know what to ask it is just a more expensive way to get mediocre results.

So that's what Bayes does. We upgrade the thinking. The name comes from Bayes' theorem — start with what you believe, test it against evidence, update your position. It's how I approach everything: strategy, enablement, writing, and occasionally arguments in the pub.

The Bayes Method

Anxiety to agency

Most people's first honest reaction to AI is fear. Will it take my job? Am I already behind? Is everything I've built about to become worthless? That fear is rational — and it's also the thing that stops people from actually learning. The shift happens when someone moves from "AI is a threat to me" to "I can do things with AI that I couldn't do alone." That's not a mindset trick or a motivational poster. It's what happens when you give someone a concrete experience of AI amplifying their own thinking — their own expertise, their own judgement — rather than replacing it.

Specify, modify, verify

Every productive interaction with AI follows the same loop. You specify what you need — clearly, precisely, with enough context that the AI can do something useful rather than something generic. You modify the output — shaping, redirecting, pushing back, asking it to go deeper on the parts that matter and ditch the parts that don't. Then you verify — checking the result against your own expertise and judgement, because the machine is confident whether it's right or wrong. That loop sounds simple. Doing it well takes practice, and the difference between someone who's practised and someone who hasn't is the difference between a useful output and an expensive autocomplete.

The Mirror

AI gives you back what you put in. Vague question, vague answer. Sharp question, sharp answer. This is why "prompt engineering" is really thinking engineering — the quality of your input reflects the quality of your thinking, and AI makes that visible in a way that's hard to ignore. People who think clearly get extraordinary results from AI. People who don't get slop. The machine is a mirror, and most people don't love what they see at first. That's where the real work starts.

Thirty minutes. No pitch. Just an honest conversation about where AI fits.

Book a call