Your AI Strategy Should Fit on One Page
A forty-slide AI strategy is a document nobody acts on. Six statements on one page changes what happens next quarter.
Erin Moore
Fractional Chief AI Officer
A useful AI strategy answers six questions on a single page: what we are trying to change, what we will fund, what we will not, how we will know it worked, who decides, and what would make us stop. Anything longer is a document that gets presented once and never governs a decision.
The six statements
1. What we are trying to change. One sentence, about the business rather than the technology. "Reduce the time from enquiry to quote" — not "leverage AI across the customer journey".
2. What we will fund this quarter. Two or three specific things, ranked. If everything is funded, nothing is prioritised — the method is in how to prioritize AI use cases.
3. What we will explicitly not do. The hardest line to write and the most useful. Naming the plausible things you are declining is what stops them reappearing every month.
4. How we will know it worked. The measure and the current baseline, both stated. Without the baseline this line is decoration — see how to measure AI ROI.
5. Who decides. A named person with budget authority, plus what they can decide alone versus what escalates.
6. What would make us stop. The condition under which each funded item ends. Agreeing it now converts a future argument into a future calculation — the point of how to kill an AI project.
Why longer documents fail
A long AI strategy is usually written to demonstrate seriousness to an audience — a board, an investor, a leadership team — rather than to direct work. It succeeds at that and then sits unused, because nobody consults forty slides when deciding whether to approve a $900 tool.
The one-page version gets consulted precisely because it is short enough to re-read in the meeting where the decision happens.
What belongs elsewhere
Governance sits in the policy and the register, not the strategy. Vendor detail sits in the evaluation. Roadmaps change; the six statements should survive a quarter.
For the language regulators and enterprise buyers use — worth borrowing so you are not translating later — the OECD's AI Principles and NIST's AI Risk Management Framework are the two most commonly referenced.
Filling it in
If you have a measured baseline and a ranked list, this is a one-hour exercise. If you do not, writing the page will reveal that, which is itself the most valuable output. The 90-Day AI Playbook walks the sequence that produces both.
Frequently asked questions
What should an AI strategy include? What you are trying to change, what you will fund, what you explicitly will not, how you will measure it against a stated baseline, who decides, and what would make you stop.
How long should an AI strategy be? One page. Longer documents are written to demonstrate seriousness rather than to direct decisions, and they are not consulted when decisions get made.
What is the hardest part to write? The list of things you will not do. It is also the most useful, because it stops declined ideas returning every month.
How often should it be revisited? Quarterly. The six statements should survive a quarter; if they change monthly, they are a roadmap rather than a strategy.
Further reading
Tagged with:
Ready to Automate Your Business?
Let's discuss how AI automation can deliver measurable ROI for your organization in 90 days or sooner.