Services
Most AI programmes stall between the workshop and the first release. Four to six weeks of structured work closes that gap: an honest readiness picture, a shortlist ranked by value and effort, and a roadmap costed against the data and team you actually have.
The blockers we get called in to unpick are almost never about models.
Every department has an idea and none of them are ranked, so effort spreads thin across all of them and nothing reaches production.
Pilots get scoped against data everyone assumes exists, and the gap surfaces mid-build, which is the most expensive moment to find it.
The strategy lands, the consultants leave, and nobody owns the next step or the budget line sitting behind it.
Review, approval, and risk ownership get designed retrospectively, which is the worst possible time to be designing them.
Four to six weeks, ending in something your team can start on Monday.
Data, platforms, skills, and appetite for change audited, then a plain statement of what you can support today rather than in principle.
Workshops surface the candidates; each is scored on value, effort, data availability, and regulatory exposure — including the ones we tell you to drop.
Phased sequencing with budget ranges and the dependencies that decide what genuinely has to come first, rather than what is most exciting.
The review model, ownership map, and skills plan that keep it moving after we leave — written so another firm could execute it.
Two representative slices of a typical engagement.
Every candidate scored against value, effort, data readiness, and regulatory exposure, so the shortlist that comes out is a decision with reasoning attached rather than a ranking of enthusiasm.
The shortlist sequenced into phases with owners, budget ranges, and the platform or data work each phase depends on before it can start.
Where AI earns its place in your business, and where it does not.
Data, platforms, skills, and appetite, assessed plainly.
Candidates ranked on value, effort, data, and risk.
Phases with budget ranges and real dependencies.
What your current foundation can actually support.
When a product already solves it, we say so.
Review, approval, and ownership designed up front.
The skills plan that outlasts the engagement.
Short on purpose — a roadmap that takes a quarter to write is stale before anyone builds.
We look at the data, platforms, and team you have rather than the target architecture, and write down what that supports today.
Candidates gathered across the business, then scored on value, effort, data availability, and regulatory exposure using criteria you agreed first.
The shortlist becomes phases with budget ranges and dependencies, including the platform work that has to land before the interesting parts can.
Owners named, the review model agreed, and the plan written so your team or another supplier can pick it up without us in the room.
Deliverables, not projections.
4–6 weeks
From kickoff to a costed roadmap, not a quarter of workshops.
One ranked shortlist
A scored set of use cases replaces an unordered wish list.
Named owners and budgets
Every phase leaves with a range and somebody accountable for it.
A readiness assessment and costed roadmap typically runs four to six weeks. It is short on purpose: a roadmap that takes a quarter to write is out of date before anyone builds against it, and the models will have moved on again by then.
It scales with how many parts of the business are in scope and how much data investigation the assessment needs, and we quote a fixed fee for a defined scope rather than an open-ended day rate. One thing worth knowing whoever you hire: the consulting fee is rarely the whole cost — platform licensing, integration work, and ongoing maintenance sit on top, and a roadmap that ignores them is not really costed.
No, and if you did you would not need the assessment. Part of the work is telling you which use cases your current data supports now and which need groundwork first. That split is usually the most useful page of the report.
We can, and often do — which is exactly why we are direct about it. We also build the seven delivery services listed on this site, so the roadmap is written to be portable: scored criteria you agreed, reasoning attached to every recommendation, and enough detail for another supplier or your own team to execute it. You are buying a plan, not a lock-in. If a recommendation only makes sense when we deliver it, we have not done the job properly.
Then that is the finding, and it goes in the report with the reasoning. Plenty of problems are better solved by an integration, a process change, or software that already exists — the no-go list is a deliverable, not an awkward conversation. It is cheaper to hear it in week three than after two quarters of build.
Yes, and often the honest answer is that what you already own is enough. Part of the review is build-versus-buy: where a product in your existing stack covers the use case, we say so rather than scoping a bespoke version of it.
A ranked use-case shortlist with value and effort estimates, a data and platform gap analysis, the no-go list with reasoning, a phased delivery plan with budget ranges and named owners, and a governance model covering review, approval, and risk ownership.
Tell us where you are stuck. We will bring the assessment, the scoring, and the roadmap.