Can AI Do Positioning? You Can Automate the Tasks, Not the Work

Positioning breaks down into a series of knowable, repeatable steps.

But if you think that means it can be automated, you're smoking bananas.

Here's why.

The process is real. It's also not the point.

Yes, positioning is a sequential process. Understand the context. Set the goals. Define the playing field. Understand the buyer. Name your advantages. Articulate your unique value. Crystallise it in a few words. Scale it across your go-to-market.

You can read the books. You can study the frameworks. You can have AI do both for you in an afternoon.

But when you actually run the process with a real company, the work turns out to be much more subtle than the tasks imply.

Positioning, like most knowledge work, is bigger than the steps you take to do it.

The tasks vs. the work

Take three steps from any positioning process and look at what actually happens in the room.

Understanding context. The task is to map the history of the company and the category — where it was, where it is, where it's going. The work is getting a room full of people with egos to agree on one story. The founder remembers a pivot one way. The head of sales remembers it another. Until they converge, there is no context to understand — only competing memories.

Naming your advantages. The task is to list the reasons customers say they buy. The work is getting at the truth underneath what they say. Customers rationalise. They tell you "features and price" when the honest answer is "your founder answered my email at 11pm and I trusted her." Stated preference and revealed preference are different animals — and only one of them is sitting in your survey data.

Articulating unique value. The task is to write a few short phrases and pick the best one. The work is getting the company to align on the bet you're making — and own the decision. A position is a refusal. It means walking away from buyers, markets, and messages that no longer fit. No sentence, however sharp, does that for you. People do. Or they don't, and the beautiful phrase dies in a document no one opens again.

Same pattern at every step. The task produces an artefact. The work produces a commitment.

This isn't a new mistake. AI just made it cheaper.

People have been productising positioning for years — selling the steps as if they were the work. Templates. Workbooks. Certification programmes. The two-day workshop with the laminated canvas.

AI magnifies the same old error. It just does it faster, and with better formatting.

And to be clear: this isn't an anti-AI position. We use AI in our own strategy work, and it genuinely helps — for research synthesis, for pressure-testing language, for generating options no one in the room would have written. The tasks get faster. That's real value.

But you can automate every task, execute the process exactly as written, and still not have positioned anything.

Because the output of positioning isn't a document. It's a decision a company actually makes — and keeps making, in pricing, hiring, roadmap, and sales conversations, long after the workshop ends. That's also why product positioning and brand positioning can't be separated from how the business itself operates.

The hard part doesn't fit in a prompt

The process is the easy part. The hard part is the judgment, the people dynamics, and the follow-through.

Judgment: knowing which of the seven plausible positions is the one this company can actually prove.

People: getting the CEO and the CRO to stop politely disagreeing and commit.

Follow-through: making sure the position shows up in the next pricing call, not just the next deck.

That's the work. And the work is why positioning done properly changes companies — while positioning done as a checklist changes the wording on a homepage.

The tasks fill a document. The work fills a decision.

P.S. Positioning is just where we happen to watch this problem play out. The same tasks-vs-work confusion is everywhere AI touches knowledge work — design, research, consulting, code. Wherever the deliverable looks like a document but is actually a decision, automation gets you the document.