AI strategy: a roadmap your CFO can fund

A board asks for an AI strategy. What comes back is either a slide deck with no engineering in it, or an engineering plan with no business case in it. Both stall for the same reason: whoever wrote them could only see one half of the problem.

Discovery closes that gap. It produces a plan that a CFO can fund and an engineer can start on Monday, because the same team wrote both halves and can defend the trade-offs between them.

It is deliberately short, and it ends in a decision.

• Leadership sessions

  What the business is actually trying to change, commercially rather than technically. Where the pressure is coming from, and what success would look like to the people funding it.

• Workflow analysis

  How the work is done today, including the spreadsheets and manual steps nobody documents. The real opportunities usually sit there.

• Data and systems feasibility

  What data exists, what state it is in, who is allowed to see it, and which systems would have to be touched. Most AI plans die here, so we find out early.

• Opportunity map

  Candidate use cases scored on value, feasibility and risk, so the sequence is defensible rather than whichever idea had the loudest sponsor.

• Governance review

  What has to be true for legal, security, and risk to sign off, established before anything is built rather than discovered at the end.

Almost every expensive AI failure traces back to skipping this. A team builds the use case that was easiest to describe, rather than the one that was most valuable, and finds out at rollout that the data was not accessible or the process owner was never consulted.

To rank your own candidates first, our guide to which processes to automate first has a scoring worksheet for value, feasibility and risk.

Discovery is the cheapest step in the programme and the one that determines whether the rest of the money is well spent.

If you want a rough read before talking to anyone, the free AI readiness assessment covers the same areas in 13 questions, and shows the result on the page.

Part of the feasibility question is which kind of AI each use case needs: a model trained on your own records, a language model, both or neither. Our guide to machine learning vs generative AI, listed at the end of this page, sets out how to tell, and which specialist each one calls for.

Sometimes build versus buy ends at an off-the-shelf assistant plan rather than a build. When it does, our comparison of ChatGPT Enterprise and Claude Enterprise sets out what each plan publishes on price, controls and data location, and what to test before signing. When the purchase is Codex for an engineering team, our explanation of how Codex pricing works shows how to estimate the monthly spend from a two-week pilot.

When it ends at a build on data that has to stay inside your own boundary, our guide to private LLM deployment compares a vendor API, a private cloud deployment and a self-hosted model on what each one keeps private.

Build versus buy also settles who should do the build, and our guide to AI agencies and AI consultancies sets out when each fits.