The Premise

There is no such thing as an AI strategy.

That sounds like provocation. It is meant to. But it is also the operating premise of this framework, and once you accept it, most of the questions that paralyze AI adoption efforts disappear.

The phrase “AI strategy” implies that AI deserves its own strategic plane — that alongside your growth strategy, your operating model, and your customer experience strategy, there is now a separate document about what your organization will do with this technology. That implication is wrong. AI is not a parallel concern. It is a set of capabilities that may or may not advance the strategy you already have. The right question is never “what is our AI strategy?” The right question is “what is our strategy, and where might AI advance it?”

This sounds like a small reframe. It is not.

When organizations start with the technology, a predictable sequence unfolds. A committee is formed. A consultant is engaged. Workshops are held. A document is produced. The document contains principles, pillars, an investment thesis, a vendor scan, and a roadmap. The roadmap has phases. The phases have target dates. By the time the document is finished, three things are true: the landscape has shifted again, no one in the organization has actually used the technology in a way that matters to the business, and the leadership team is now committed to a vision they have no evidence for. Months pass. The deck gets revised. Nothing ships.

You have probably seen this happen. You may have sat in the rooms where it happened. The instinct that produced the deck is not wrong — leaders are right to want clarity before they commit resources. The mistake is the order. Clarity, in this domain, comes from doing, not from planning. The technology moves too fast and behaves too strangely for advance specification to hold up. What you need before you commit resources is not a comprehensive plan. It is a small number of running projects whose results you can read.

If you flip the order, the work changes shape. You begin with the business — what it is, what makes it work, where it is going, and what threatens it. You become specific about the operations that produce value and the moments where customers experience that value. You name the capabilities that matter. Only then do you ask, in concrete terms, where AI might advance any of this. The conversation moves from the abstract to the particular. It becomes possible to evaluate ideas against something. It becomes possible to say no.

That last part matters more than it sounds. A framework worth using has to leave room for the answer “AI does not make sense here, at least not yet.” When you start with the technology, that answer is structurally hard to reach — the whole apparatus of the engagement is built to produce a yes. When you start with the business, no is a perfectly normal output. Some parts of the company will turn out to have AI-shaped problems. Others will not. A strategy that cannot conclude “not here” is not a strategy. It is a procurement plan.

There is a subtler failure mode worth naming. Even organizations that nominally start with the business often slip into a hybrid posture — they describe the strategy in broad terms, then move quickly to a list of “AI use cases” disconnected from any specific operational reality. The list looks like progress. It is not. A use case that is not anchored in the actual business is a wish, not a project. The discipline this framework asks for is the discipline of staying anchored: every candidate AI initiative must pass through the question, “what part of our strategy does this serve, and how will we know if it does?”

This is why the rest of this document begins where it does. The first work with any organization adopting AI is strategy work — not AI strategy work, just strategy work. Mapping the business model. Articulating what you are trying to become. Identifying where the friction lives and where the leverage is. AI enters the picture only after that map is drawn, and only as a candidate answer to questions the strategy has already framed.

If you do this in the other order, you will spend months producing a document that looks like progress and produces none. If you do it in this order, you may be running your first projects in weeks.

That is the whole premise. Everything else in this framework follows from it.