AI did not come for the strategist’s job. It has changed what the job is.
The list of things that make someone a good strategist is being rewritten. Some of what is on it today will matter less in a few years. Research speed. Deck polish. The volume of starting ideas you can put on a wall in a room. New things will appear, and one in particular is doing most of the work: the ability to build your own AI tools. The strategists who can do that will be the ones who reach the new ceiling first.
A few weeks ago, while working on a pitch, I watched a tool I had built do several stages of strategic planning in a few minutes, work I have spent my whole career doing by hand. I felt something. Not about the time saved. About the fact that something I had shaped was now lifting the work it was made for. That is the move.
Strategy uses both hands
Strategy has always been an unusual discipline. It asks for analytical work and creative work from the same person, often inside the same hour. At Brainlabs we sometimes talk about it as a left-brain, right-brain craft, and the metaphor holds up well enough for what it is doing here. The analytical half sizes markets, builds business cases, and holds the structure of an argument together. The creative half reads what people actually want, what they will not say in a focus group, and develops ideas against a brief grounded in that understanding.
What is different about this wave of AI is that it is genuinely good on both sides. It finds patterns in data and pulls signal out of a body of research. It also brainstorms, finds odd angles, and supports synthetic audiences to pressure-test a hypothesis against before anything goes to market. One tool, doing both, at once.
The floor is mastery. The ceiling is what you build.
Mastering AI puts a strategist at the floor, and the floor is rising. The things that used to mark a strategist out are turning into table stakes.
In a recent letter to the Financial Times, Neil Lawrence, the DeepMind Professor of Machine Learning at Cambridge, made the macro version of the case. AI value, he argues, will not arrive in one lump to be distributed from the top. It comes “workflow by workflow, in the hands of the people who understand the work.” Google’s Generative AI Leader course frames the same shift operationally. A real AI strategy works in two directions. Leadership sets priorities from the top. The people doing the work surface what they need from the bottom. Most organisations are doing the top half well and the bottom half barely.
For strategists, the bottom-up half has a shape, and it is tool-building. I used to ask what a tool could do for my strategy. Now I ask what my strategy needs, and I shape the tool around the answer. A tool built around how the work actually gets made (the questions it asks, the inputs it depends on, the form of its outputs) lifts every piece of work that runs through it in ways an off-the-shelf product never will. That reversal is the new ceiling. The old one was the limit of what someone else’s product could do for the work. The new one is the limit of what you yourself can shape into a tool. The line is far higher.
What it has taken
None of this is a free upgrade. For me it has meant structured learning across Google’s Generative AI Leader course, the Anthropic Academy, Notion Academy, and more. Brainlabs gives everyone protected time on Innovation Tuesday, which is when the building actually happens. Several of our builds are now running inside live client work, each one shaped around a specific part of the planning process: generating insight, automating research, surfacing best-in-class case studies, supporting creative ideation. When a prototype proves out and needs to scale, engineering partners from the central tech team step in. The mature builds consolidate over time on our shared internal platform, Cortex.
Three moves for brand-side leaders
If you are running a brand-side team, three moves are worth making now.
Invest in capability, not just access. Real capability lives in understanding, and understanding moves with the field. What your team learnt about a model six months ago is already partly out of date. They need ongoing exposure to what the current generation can do, time across more than one platform, and a working grasp of what sits underneath the products: context windows, the trade-offs between model families, when to reach for which tool for which job. Structured learning produces that fluency. Licences on their own do not.
Map the workflow before you build. Every team has its own way of operating. Before any tool gets built, the team needs to be clear on what AI should take over, where it should support, and where a human has to make the call. That mapping is the bottom-up half of an AI strategy in practice, and it is what serious tool-building rests on.
Put a support system around the building. Protected time has to be genuinely protected, or nothing gets built. Engineering partners have to be reachable for the moment a prototype proves out and needs to scale. Without that scaffolding, only the most stubborn strategists ever reach the ceiling.
Do these three things and the upside is concrete. You get a different kind of strategy work, and strategists who can take a team somewhere it has not been before.
Building your own tools used to be a side project. It is the job now.
