strategyplaybook.aiSqwyz
Sign in
← All 124 tools
Execution · Plan

AI operating model options

Chooses how to structure AI ownership — centralised, federated, hub-and-spoke or embedded — against the organisation's size, maturity and portfolio, not against what looks impressive.

The printable canvas and the handoff into Claude Code are part of a paid plan. See what a plan includes. It needs the Claude desktop app on this machine, and your team and the prompt library already installed in that project — we cannot see your disk, so open it there.

Stage
04 Plan
Works at
Corporate, Business
Maturity
Scaling → Large organisation
Time
Three to five days

What it is

Compares four structural options for owning and running AI capability as it moves from experiment to production, tested against the organisation's actual size, maturity and portfolio shape rather than against whichever model looks most sophisticated on paper.

Centralised
A single team owns AI capability and delivers it as a service to the rest of the organisation. Fastest to build genuine expertise; slowest to embed AI into the specific judgement each function needs.
Federated
Each function builds and owns its own AI capability independently. Fastest to embed in specific judgement; slowest to build deep expertise and likely to duplicate effort across functions.
Hub-and-spoke
A central team sets standards, tooling and governance while embedded specialists in each function do the applied work. The most common landing point for organisations past the earliest stage, and the most complex to run well.
Fully embedded
AI capability dissolved entirely into existing functional teams with no central coordination at all. Rare, and usually only right for organisations where AI has become simply part of how every function already works.
The fit test
Chosen against the organisation's actual size, maturity and how concentrated or spread the AI portfolio is — not against which model is currently fashionable.
Revisit trigger
A stated point — a maturity threshold, a portfolio size, a specific initiative — at which the model should be reconsidered, because the right model at experiment stage is rarely the right one at scale.
The mistake people makeChoosing the model that fits where the organisation aspires to be rather than where it actually is now, which produces a structure with more governance than the current AI portfolio can justify.
What it’s forAI capability is moving from experiment to production and nobody has explicitly decided who owns it structurally.
What it’s not forAI activity is still purely experimental with no production capability yet. An operating model decision made before there is anything to operate tends to fit the org chart rather than the actual work.

How you run it

  1. Map current AI activity and who runs itEvery initiative in flight or in production, and its current ownership — the honest starting point rather than the aspiration.
  2. Score the organisation's size and maturityAgainst each of the four models' actual requirements — a hub-and-spoke model needs a scale of AI activity a small federated effort does not yet have.
  3. Compare the four models against that specific fitNot against which model looks most sophisticated. The right model for a twenty-person organisation is frequently not the right model for a group with a dozen business units.
  4. Choose and name the model, stating the trade-off acceptedEvery model trades speed against depth, or coordination against duplication. State explicitly what was given up.
  5. Set a revisit triggerA maturity threshold or portfolio size at which the choice should be reconsidered — because the right model at experiment stage is rarely the right one at scale.

The prompt

Two ways to run it

Run this tool in your own Claude

The short prompt starts your partner against the library on your disk. The long one carries everything with it and needs nothing installed.

🔒 Prompt locked

Everything above is free — what this tool is, what it is for, what it is not for, and how to run it as a workshop. Three tool prompts a month are free with an account; beyond three, and for the printable canvas, it is the paid part.

See plans — from £19 a month

Your playbook

It lands in the earliest stage this tool suits. Move it on the Playbook page.

You’ll need

  • The current AI initiatives in flight or in production, and who currently runs each
  • The organisation's size and existing operating model conventions for other capabilities
  • How the AI portfolio is expected to grow over the next twelve to eighteen months

You’ll end up with

  • The operating model that fits the organisation as it is now, chosen and named
  • The trade-offs of that choice stated explicitly, not glossed over
  • A stated trigger for when the model should be revisited

Runs well with

AI risk classification and governanceAlliance and partnership designAssumption logBackcasting