Another AI idea will not tell you which one is worth funding.

Every department has a list by now. Product wants an agent for one thing. Support wants to automate another. Someone in finance saw a demo and wants to try a third. Each idea sounds reasonable on its own, and none of them is obviously wrong.

The problem shows up when you try to rank them. Reasonable people disagree, because nobody has connected the list to the parts of the business that actually matter: which value stream carries the most profitability, which one is the constraint on growth, which one is eating the most labor, and where AI is genuinely ready to help versus where it just sounds good in a slide.

Without that connection, prioritization becomes whoever argues loudest, or whoever saw the most recent demo. That is not a technology gap. That is a focus gap.

The Sprint starts with your business, not with AI. We map the value streams that actually run your company, and score the candidates against a small set of factors: economic impact, whether it's a constraint on growth, how labor-intensive it is today, and how ready it actually is for AI right now versus later.

I use something WSJF-shaped for this: not because the math is precise, it isn't, but because ranking opportunities against the same factors, out loud, in the room, is what turns an argument into a decision. What you leave with is a prioritized list of your own: a treasure map of your AI opportunity, not mine.

You can get a list of AI use cases from a vendor demo, a conference talk, or an afternoon with an AI tool. That isn't what's missing. What's missing is a defensible way to decide which of those use cases is actually worth your team's time and your budget, grounded in your business rather than in whatever AI happens to be good at this month.

A generic AI strategy engagement usually ends the same way: a list of things to build. The list is rarely the gap. The gap is who owns the decision, what evidence would justify the investment, and whether anyone can tell six months from now if it worked. This Sprint is built to produce that, not another list.

For the handful of opportunities that survive prioritization, we go deeper. For each one, we get specific about what "amazing" would actually look like: who does what, by how much, and how you'd know it happened. That's the Opportunity Canvas: your hypothesis written down, not a slide promising transformation.

Then we map how work actually flows through that area today: where it waits, where it's stuck, what's still manual. From there we sketch what the target state could look like, as a real progression: human-led now, AI as a copilot, AI running with a person checking its work, AI running with a person watching from a distance. Not every opportunity needs to go all the way down that path, and part of the work is being honest about which ones do.

We close by sorting the finalists on opportunity against risk, so you leave knowing not just what to pursue, but what you'd need to learn first to de-risk it.

This tends to fit leaders at companies roughly 100 to 1,000 people: big enough that the coordination problem around AI investment is real, small enough that a leadership team can still act on what it decides.

This is worth looking at when:

• You have more AI ideas than a way to rank them: Pilots, licenses, and a growing backlog of "we should try this," with no shared way to say what matters most.
• You can't measure what AI has actually done: Real money spent on tools and consultants, and the honest answer to "what changed" is unclear.
• Activity keeps not turning into outcomes: Things ship, demos land well, and the business still does not feel any different.
• You're not fully sure who should run this: Not sure your own team has done this kind of prioritization before, and wary of outside help too, because very few people actually have.

One more thing worth sorting out early: whether your company is genuinely product-intensive, building and shipping something of your own, or more services-intensive without much in-house product capability yet. That changes where the real opportunity actually sits, and it's worth being honest about rather than assuming.

I don't do vendor or tool selection. If the Sprint surfaces that you need a particular platform or vendor, that's a separate decision for you and your team to make, not something I sell into.

I also don't build it for you. If an opportunity is worth pursuing, the point is to guide your team through building the capability themselves, close enough to the work to catalyze it, not to hand you a finished thing you now depend on me to maintain.

That's also the honest answer to the trust question underneath most of this: if you don't fully trust your own team to run this kind of prioritization alone, and you're wary of outside help too because so few people have actually done it, the answer isn't picking one or the other. It's working through it together, in the room, until your team can run the next one without me.

You leave with a prioritized view of your AI opportunity: your own version of the treasure map, showing which value streams matter most, and why.

For the two or three opportunities that rose to the top, you get an Opportunity Canvas for each: what "amazing" looks like, who does what by how much, and what would tell you if it's working. You get a picture of how work actually flows through those areas today, and a realistic target state, not a wish list. And you get those finalists sorted by opportunity against risk, so you know what to try first and what you'd need to learn to de-risk it.

What you don't get here is a price or a delivery schedule. Those depend on what we find, and they're a conversation once we both know what's actually worth pursuing.

Bring whatever you've already got: the pilots, the ideas, the pressure from the board to "have an AI strategy." No pitch, no assumption that AI is the answer everywhere. We'll talk through where the real constraint is and whether this Sprint is the right next step.