Services

AI strategy and advisory

Set the AI agenda for the whole institution: where AI creates advantage, where it creates exposure, and the sequence to get there. Then I lead it as a fractional Chief AI Officer until it is real, not a slide.

The problem it solves

Most institutions do not have an AI problem; they have an AI drift problem. Pilots multiply, vendors circle, the board hears "we're doing AI," and yet nothing changes how the business actually runs. The strategy deck gets approved and quietly filed. What is missing is not ambition, it is a direction tied to real outcomes, an operating model that survives contact with the organization, and someone accountable for making it real. This work supplies all three, and stays until AI is working in the business rather than sitting in a report.

Who it is for

CEOs, boards, and executive teams who need an AI direction they can put budget and reputation behind, and someone senior who will stay to make it happen, not hand over a slide deck and leave.

What you get

• An AI strategy and roadmap tied to real business outcomes, not technology for its own sake
• The operating model and governance structure to run AI across the institution
• Build-versus-buy and vendor decisions made on evidence
• Fractional Chief AI Officer leadership through execution
• Board-ready briefings that turn AI into a decision leadership can own

How it works

1. Fix the real objective before anything is committed
2. Set the strategy and the roadmap
3. Stand up the operating model and governance
4. Stay through execution until it is working

Questions

Common questions

Our board keeps asking what our AI strategy is and honestly we do not have one. Where do we start?

Start with the outcome, not the technology. In the first working session we name the two or three decisions or workflows where AI would move a number the board already tracks, then test whether your data and your people can support that. Everything else is sequencing. A usable strategy fits on a few pages: where you will apply AI first, what you will not do this year, who owns it, what evidence comes back to the board, and what has to be in place before you spend.

What is the difference between people using ChatGPT at work and an AI strategy the board can sign off on?

Individual tool use is productivity. A strategy is a set of choices the institution can be held to. Which decisions and workflows AI touches this year, which it does not, what data it may use, who owns the outcome, and what evidence returns to the board. Without those choices you get scattered pilots, duplicated spend, and no way to tell whether anything worked. With them, tool use becomes an asset instead of an unmanaged risk sitting outside anyone's view.

How do we decide which problem AI should solve first instead of trying everything at once?

Pick the one where a number you already report moves, the data exists, and a wrong answer is survivable. That combination is rare enough that it usually settles the argument in a single session. I open engagements this way deliberately: name the outcome, test feasibility against your real data and workflow, then sequence everything else behind it. Running five pilots at once mostly buys you five inconclusive results and a tired team that has stopped believing the program.

Is it realistic to pilot AI in one product line first before rolling it out across the business?

Yes, and it is usually the right call, provided the pilot is designed to settle a question rather than to impress. Choose one line with clean enough data and a measurable outcome, define in advance what result would make you stop, and run it long enough to see the failure modes. Then decide. The common mistake is piloting in the easiest corner of the business and generalizing from a result that never had to survive real volume.

Related: The Aperture method · How a CEO should approach AI · Selected work · All services