Setting the North Star

Before AI enters the conversation, the strategy has to. The first work with any organization adopting AI is not an AI workshop, an opportunity scan, or a tooling assessment. It is focused strategy work whose output is a shared description of the business — clear enough that AI candidates can later be evaluated against it. Until that description exists, no AI conversation will produce useful answers, because nobody yet has a stable frame for what “useful” means.

This chapter is about producing that description. It populates the Strategic Loop — the top layer of the architecture from Chapter 3. The loop will keep evolving as projects run and evidence accumulates, but it has to start somewhere. This is where it starts.

The four anchor questions

The strategic work organizes around four questions. They are deceptively simple, and the leadership team usually does not agree on the answers right away — which is itself the first useful output of the work.

What is the business? A working description of the model itself: what you sell, who you sell it to, how the money flows, what the key activities and resources are. Not a mission statement. A description that, read by an outsider, would let them recognize what the business does and how it makes money.

What makes it work? What is genuinely distinctive about this organization — the capability, relationship, asset, or dynamic that makes it competitive? This is the question most often answered with a comfortable cliché (“our people,” “our culture,” “our customer focus”) that does not survive five minutes of follow-up. The real answer is usually more specific and less flattering, and finding it is part of the work.

Where is it going? A clear statement of what the organization is trying to become over the next three to five years. Not aspirational language. A picture of the future business specific enough to commit to and concrete enough to be wrong about. If the answer is genuinely “we don’t know yet,” that is itself a useful answer — it means the strategic work is not done, and an AI adoption effort built on top of it would be premature.

What threatens it? The actual risks to the business — external pressures, internal vulnerabilities, structural shifts, emerging competitors, capability gaps. The honest version of this question is hard to answer in a room of executives because nobody wants to be the person who names the unfixed weakness. The work here is to make that conversation safe enough to have.

These four answers together constitute the North Star. Everything that follows in the framework is evaluated against them.

What the work actually looks like

The leadership team is in the room, working through the four anchor questions in facilitated conversation. The answers belong to the team; the facilitation surfaces them. The output is a set of working documents, not a polished deliverable. The North Star is a tool, not a poster.

By the end of the strategic work, you have a set of artifacts — a business model description, an articulation of the competitive engine, a statement of strategic intent, and a risk landscape — that the team has agreed on with enough confidence to act on. They are not final. They are good enough to evaluate AI candidates against, which is the only thing this stage of the work needs to produce.

Time is the warning sign here. If this work is dragging on, something is wrong: either the team cannot agree on what the business is (in which case the AI adoption effort is the wrong problem to be solving), or the work is being over-polished into deliverables that look impressive but no longer help anyone make decisions.

What to keep out of the room

Three things, in particular, should not appear in these initial sessions.

The first is AI itself. The whole point of starting here is to know the strategy before asking where AI fits. If the conversation drifts into “what AI can do,” the discipline is broken before it has had a chance to work. There will be plenty of time for the AI conversation. It begins in the next chapter, after this work is done.

The second is the pre-existing opportunity list. There is a strong temptation, especially in organizations that have been thinking about AI for a while, to come in with a list of AI use cases and try to map the strategy onto them. This inverts the order. The strategy comes first; the opportunity list is generated from the strategy, not the other way around.

The third is over-formalization. The North Star is not a board document. It is a living set of working artifacts that the team will revisit as projects run and reality teaches them things they did not yet know. If the document is being treated as final or precious, it has stopped being useful. The right tone is “this is what we currently believe, and we expect parts of it to be wrong.”

When to move on

The strategic work is complete enough when the leadership team can evaluate a proposed AI initiative against the four anchor questions and reach a shared judgment about whether it serves the strategy. If three executives can read a candidate project description and arrive at meaningfully different conclusions about whether it advances the business, the strategy is not yet specific enough. If they converge — even after reasonable debate — on a shared answer, the North Star is doing its job.

That is the threshold for moving on. With the strategy in place — the sunlight everything below it grows toward — the question becomes where, specifically, AI might advance any of it.