Home What we do AI strategy & diagnostic

Ship Shift · Capabilities

AI strategy & diagnostic

Last updated: 2026-11-17

Decide which AI bets to fund — and which to park — with a signed roadmap, not a trendy slide deck.

Strategy and planning session in an office

The problem: spending without a return criterion

The risk isn't «not doing AI». It's funding three demos at once, each with a different vendor, with no shared view of volume, effort and risk. Leadership sees slides; operations keeps typing. A year later there's no metric that justifies the next investment.

A typical committee piles vendor demos, ownerless PoCs and an «AI ideas» backlog nobody ranks with numbers. The cost isn't just the licence — it's operations attention diverted into experiments without success criteria. The signed roadmap exists to stop that focus leak.

When we prioritise bets we ask for volume, effort and risk evidence. Without those three variables any ranking is opinion.

What we don't do in the diagnostic

No generic internet use-case catalogue. No miraculous ROI without sampling your flows. Close is a fixed-price pilot proposal with signable acceptance.

We also refuse infinite audits. Three well-used weeks are enough to know what to fund first. If sampling shows the winning bet is integrating two systems — not a model — we say so, even if that shrinks the «AI» scope of the next contract.

Without a signed map, teams pick the most visible case — or the one the loudest vendor pushes — not the one that unlocks real hours in the process that hurts most.

Our method: prioritise bets, not tools

We start with the people who live the process. In a few sessions we measure volumes, exceptions, systems of record and human dependencies. We rank each candidate by expected return, integration cost and operational risk. You get a ranked list of bets: fund first, backlog, or discard.

  1. Interviews and observation on the real flow (not a generic survey).
  2. Inventory of systems, data and friction points.
  3. Prioritisation matrix: volume × effort × risk × ownership.
  4. Roadmap with signed success criteria and a scoped pilot.

Deliverables

  • Process and use-case map with owners and volumes.
  • Ranked AI bets with return rationale.
  • Quarterly roadmap with indicative costs and dependencies.
  • First-pilot scope with fixed price and acceptance criteria.
  • Responsible-AI notes and AI Act fit where relevant.

Who it's for

  • Leadership and operations deciding AI budget without blind bets.
  • Digital transformation teams that need an actionable map.
  • Mid-sized companies with repetitive processes and pressure to show ROI.

Risks and compliance

From day one we flag where human supervision, decision logs and personal-data limits are required. We don't sell «AI everywhere» — we sell an investment order you can defend in committee and audit.

What the engagement looks like

01

Week 1

Kick-off, access and stakeholder map.

02

Weeks 2–3

Workshops, volume sampling and draft ranking.

03

Close

Signed roadmap + fixed-price pilot proposal.

FAQ

Is the diagnostic a fixed fee?
Yes — we scope it with you. Usually a closed 2–3 week engagement with concrete deliverables.
Do we need clean data to start?
No. Part of the work is seeing what data exists, where it lives and what to fix before the pilot.
Will you stay to build afterwards?
Yes. The diagnostic ends in a pilot proposal. We don't sell orphan slide decks.
How is this different from a big-consultancy RFP?
We work on the real flow and leave with a fundable pilot — fewer abstract frameworks, more ranked bets.
Do you cover public funding guidance?
When it fits, we point to relevant schemes and evidence required — without promising awards.

Want a ranked list of bets — not another demo?

Tell us your industry and the process that hurts most. On a call we'll say if a diagnostic makes sense.

Let's talk