In November 2025 I had built a personal AI agent, and one of the largest wealth management firms in East Africa reached out. They were in a growth crisis. They needed their people to be more productive, and they thought agents were the answer. I went in confident — this was exactly what I had built.
The first session went exactly as expected
We sat down to gather requirements. Agents that could handle research, summarise documents, draft communications. Standard personal-productivity work. I was confirming the obvious and feeling good about it.
Two hours into the second session, it shifted
The employees weren't actually showing me their personal productivity problems. They were showing me their handoff problems. The moment value gets lost in their firm isn't when someone is doing their work — it's when they pass it to the next person. Email flattens the context. The person who knows what happened is often unavailable. Everyone downstream starts from scratch.
And they had a hard constraint that made it acute: they can't hire experts from Europe at Western salaries, and the local mentoring infrastructure to grow that expertise quickly isn't there. They couldn't just throw money at it. What they needed was a way to package a unit of expertise, complete it, and hand it off — without a human babysitting every transfer.
A personal agent is useful. Agents that can coordinate are a different category of value entirely.
The third session confirmed it
Same pattern, again. The coordination problem isn't unique to this firm or this geography — it's just most visible where the constraints are hardest. It is everywhere. Email buries context, chat buries it faster, and every handoff resets the clock.
What this became
The handoff between people is where value leaks in almost every organisation. Agents that can hand off context — verified, intact, and private — solve something much larger than personal productivity. But for agents to hand off to each other across organisations, they need what people have always needed: identity, a way to gauge trust, and a record of what was agreed.
That is what became CELLO. The discovery sessions weren't market research. They were the product design. The employees of that firm are the reason CELLO exists.