what if AI could actually coach
like a master coach?
Early in my career I worked at the intersection of language and technology — machine translation, localisation, and the early tools that were trying to make language work like data. I wasn't a developer. I was the person between the clients and the engineers — understanding what people needed, testing what was built, and making sure the technology actually worked for the humans using it. Language as data. Patterns as logic. That world shaped how I think.
When I moved into coaching I brought all of that with me — the systems thinking, the linguistic instinct, and eventually 10,000 hours working with executives, founders, managers and sales teams. I built Coaching Skills for Managers as a live programme long before it became a product category. The technology background never left. It just found a new application.
And I built a methodology. Rooted in NLP, Neuro-Semantics, and psycholinguistics. Not frameworks — a genuine decision system for reading how people think, what's blocking them, and what intervention creates real change. The kind of work that goes deep fast and holds.
The whole time I had one persistent frustration. All of it lived in the room. When the session ended, the intelligence went with it. The organisations I worked with were investing in coaching with almost no way to see what it was doing across their people. And I could only reach so many.
So I went out on a limb — because I wanted to see how far I could take things. I started mapping my own decision logic. Not what I know — when I do what, and why. The questions I ask under pressure. How I read resistance. How I track a person's language patterns across six sessions and find the lever that shifts everything.
Coaching isn't sequential. It's decision logic. The question isn't what do I know — it's when do I do what, and why. That's what needed encoding. So I rebuilt it. With Claude as my thinking partner and engineer. Forty coaching sessions to show it how I work. Six refined databases where there had once been sixty-eight.
My linguistics background had always told me that language is mathematical — patterns, weights, signals, probabilities. Then I came across a mathematical framework that intrigued me — I didn't fully understand it, and that's exactly why it caught my attention. I'm not a mathematician. But I asked Claude how we could apply it to CaaSy. We added it as a component of the system. And it improved the product no end. That was the bridge I hadn't known I was looking for. The way someone speaks tells you not just what they think, but how likely they are to change, and what kind of intervention will land.
Then I came across a mathematical framework that intrigued me — I didn't fully understand it, and that's exactly why it caught my attention. I'm not a mathematician. But I asked Claude how we could apply it to CaaSy. We added it as a component of the system. And it improved the product no end.
One agent listens. One decides the strategy. One speaks in the coaching voice. One tracks the live session. One holds the memory across sessions. One maps the organisational patterns. Each one brilliant at its single job. That architecture — six specialists running together — is what turned a smart experiment into a coaching system that actually works.
That's when CaaSy stopped being a coaching tool and became something I hadn't planned for — an intelligence platform. The coaching was always the point. The organisational data was the gift hiding inside it.
One question.
This is what happened.