The usual case for t-shaped skills is made in terms of optionality: the more domains you know something about, the more flexible you are, the more roles are available to you. This is true but not the point. Optionality is a career benefit. The structural advantage is something else.
The structural advantage is position.
In any complex organisation, most of the interesting problems — and most of the value — exist at the boundary between domains. The tension between what engineering builds and what users actually want. The gap between what a data model says and what a business leader can act on. These are not problems that sit inside one domain. They live at the edges, where two disciplines meet and neither owns the resolution.
A specialist is positioned inside one domain. They are expert at its core but often blind at its boundary. A T-shaped professional is positioned at the boundary — able to see the problem from both sides, speak both languages, and ask the question that neither side had thought to ask.
That position is the advantage. Not knowing more things. Being where the collisions happen.
I have spent all of my career at one boundary or another. Not always by design, often at organisations that were changing faster than their structures could accommodate. A digital product role that expanded into business transformation — at an organisation that had succeeded in B2B and had to learn, for the first time, how to talk to its customers directly. An engineering remit that pulled in enough commercial context that the distinction started to blur. My first stretch into education, that forced me to understand learning differently from how any product person would by default. A cross-functional scope that reached into marketing where, for once, the value flowed the other direction: the technology background was what marketing needed to go digital-first, before the rest of the industry had worked out that it was possible.
None of these adjacencies were comfortable to enter. Each one had a period of early uselessness — present in the room but not yet capable of contributing much. The transition from observer to participant took a little time, but not as much as I may have expected, and it always involved a specific kind of vulnerability, admitting that prior experience did not prepare me for the role.
What came out the other side was not just a broader skill profile: it was a different quality of outcome. The contributions I make now at the boundary between product and engineering are not ones that either a pure product person or a pure engineer would make. They are outcomes that are only visible from having operated, with real stakes, on both sides. I cannot point to a specific course where I learned to think like that. I learned by being in the teams and initiatives where the need arose, and being the person who had to do something about it.
AI can search across domains. Given a problem at the boundary between product strategy and data architecture, a capable model will surface relevant frameworks and generate a plausible synthesis. It looks like cross-domain thinking.
The difference is stakes.
The model has not been responsible for a decision at that boundary and been wrong. It does not have the scars that come from getting cross-domain judgment wrong when it actually matters. It does not bring the specific context of this organisation, this market, these relationships — the kind that only accumulates through experience you cannot shortcut.
The T-shaped professional at the boundary does have that. The model cannot replace it. What the model can do is make the T-shaped professional faster, better-informed, and less constrained by the limits of their own memory. The advantage of the T-shape does not disappear with AI. It becomes the thing the AI amplifies.
In Kasparov’s Advanced Chess experiments, the pairs that beat both grandmasters and supercomputers were not necessarily the strongest individual players. They were the players who had developed the best process for working with the machine: knowing when to trust it, when to override it, when to push it in a direction it was not naturally exploring.
That process is a cross-domain skill. You need enough depth to evaluate the machine’s output in your primary area. You need enough breadth to recognise when the machine is optimising for the wrong objective because it lacks context from an adjacent domain. The person who has only depth can evaluate but not redirect. The person who has only breadth can redirect but not evaluate. The T-shaped professional can do both.
The T-shape is not the answer to the AI question. It is the structural profile of the person who can work effectively with AI as a tool rather than being replaced by it as a function.
The next post — The T-Shaped Organisation (11 June) — scales this logic from the individual to the organisation. The same dynamic that makes the T-shape valuable for a person makes it a structural advantage for a team — and organisations that are built around it are better positioned for the next few years than those that are not.
Leave a Reply