AI Strategy Consulting
Deciding where AI actually belongs in your business — and where it does not — before anyone writes a proposal or signs a licence
Most organisations approaching AI have the same two problems: too many possible use cases and no way to rank them, or a board expecting an answer to a question nobody has framed properly. AI strategy consulting fixes the sequence. We assess what your business actually does, identify where a language model would genuinely change the economics, rank those opportunities by return and feasibility, and produce a plan you can fund. We are vendor-neutral and we sell no licences — so "do nothing here" is an available conclusion, and we have reached it.
In short
- Use cases ranked by return, not by novelty. A scored shortlist with the cost, effort and expected benefit made explicit for each.
- Data readiness assessed honestly. Most AI projects fail on data quality and access, not on models. We find that out before you commit budget.
- Build, buy or wait — a clear recommendation for each opportunity, including the ones where an off-the-shelf tool beats anything custom.
- Governance and policy — an acceptable-use policy for staff, data-handling rules, and the approval boundaries that keep you out of trouble.
- Vendor-neutral. We resell nothing and take no commission. If the right answer is a product we do not build, we will name it.
Opportunity Assessment
A structured review of your operations to find where AI changes the economics — and to eliminate the places where it only adds cost and risk.
A Ranked Roadmap
Use cases scored on business value, technical feasibility, data readiness and risk, so funding decisions have something defensible behind them.
Data Readiness Review
What data you hold, what state it is in, and what has to change before any of it can support an AI system. Usually the real bottleneck.
Policy and Governance
Staff acceptable-use policy, data-handling rules, approval thresholds and audit requirements — written for your actual regulatory context.
What the Engagement Covers
We weight these to your situation. A company with no AI in use needs a different engagement from one with six departments running unmanaged tools.
| Workstream | What we produce |
|---|---|
| Current-state assessment | What your teams already use, officially and unofficially, what it costs, and where shadow AI usage is creating data exposure nobody has assessed. |
| Use-case identification | A structured sweep of your operations producing a long list of candidate applications, each traced to a specific process and its volume. |
| Scoring and prioritisation | Every candidate scored on value, feasibility, data readiness and risk, producing a ranked shortlist with the reasoning visible. |
| Data readiness audit | What data exists, where it lives, its quality and accessibility, and the specific remediation needed before it can support the shortlisted use cases. |
| Build-versus-buy analysis | For each priority use case: an existing product, a custom build, or a hybrid — with real costs on both sides, including the ones vendors leave out. |
| Model and platform selection | Which models suit which task, at what cost, with what data-residency implications — and where an open-weight model on your own infrastructure is the right call. |
| Cost modelling | Build cost, running cost at projected volume, and the point at which usage-based pricing stops being cheaper than an alternative. |
| Governance framework | Acceptable-use policy, data classification rules, human-approval thresholds, and the audit trail your sector requires. |
| Implementation roadmap | A sequenced plan with dependencies, a pilot recommendation, success criteria, and the decision gates at which you stop or continue. |
We are a technology company that builds AI systems, and we consult on AI strategy — those two facts create an obvious incentive, so we state it plainly. Our consulting fee does not depend on you commissioning a build, our recommendations are written to be executed by any competent supplier, and "the answer here is not AI" is a conclusion we deliver regularly.
Who This Is For
Boards Under Pressure
Leadership is being asked what the AI strategy is, and the honest answer is that nobody has assessed the question properly yet.
Too Many Ideas
Every department has a proposal, the budget covers two of them, and there is no shared basis for deciding which two.
Ungoverned Adoption
Staff are already pasting company data into public AI tools, and there is no policy, no visibility and no assessment of the exposure.
A Pilot That Stalled
Something was built, it demonstrated well, and it never reached production — usually a data, integration or ownership problem rather than a model problem.
What You Receive
- An assessment report — current state, shadow AI exposure, and where your operations stand against comparable organisations.
- A scored use-case shortlist — ranked, with value, cost, effort and risk stated for each, ready to take into a budget discussion.
- A data readiness report — what has to be fixed, in what order, and what it blocks until it is.
- Build-versus-buy recommendations — named products where buying wins, scoped specifications where building wins.
- A governance pack — acceptable-use policy, data-handling rules and approval thresholds, in a form you can circulate to staff.
- An implementation roadmap — sequenced, with a recommended first pilot, its success criteria, and the gate at which you decide to continue.
- An executive presentation — a version you can put in front of a board without rewriting it first.
How We Work Through It
- Kick-off and objectives — what leadership is actually trying to achieve, what constraints are non-negotiable, and what has already been tried.
- Interviews and observation — time with the people doing the work in each function, because the best use cases are found on the floor rather than in the org chart.
- Data and systems review — what you hold, where it lives, what state it is in, and what your architecture will and will not permit.
- Analysis and scoring — candidate use cases assessed against a consistent framework, so the ranking survives challenge from any department that dislikes it.
- Prioritisation workshop — a working session with your leadership to agree the shortlist and the sequence, rather than us presenting conclusions you had no part in.
- Roadmap and handover — the final report, the governance pack, the board presentation, and a walkthrough with whoever will own delivery.
Frequently Asked Questions
You build AI systems — how can your advice be neutral?
It is a fair challenge and we answer it structurally rather than with reassurance. The consulting fee is not contingent on a build following. Deliverables are written so any competent supplier can execute them. We recommend off-the-shelf products by name where buying beats building, and we recommend doing nothing where nothing is warranted. You should still weigh our advice knowing what we sell — that is sound procurement practice, not distrust.
Is our organisation too small for this?
Possibly, and we will tell you. A small business usually needs one or two specific automations, not a strategy engagement — in that case we will say so on the first call and point you at the practical option. This work earns its cost when there are competing priorities, real data complexity, or a governance gap that needs closing.
What if the conclusion is that we should not adopt AI?
Then that is the deliverable, with the reasoning and the numbers behind it. It is a legitimate and reasonably common outcome — particularly where data is not in a usable state, or where volumes are too low for automation to pay. Knowing that costs far less than discovering it after two failed pilots.
Do you help with implementation afterwards?
We can, and often do. It is not a condition of the engagement. The roadmap is deliberately written to be executable by your internal team, by us, or by a third party — and we will tell you honestly where we are not the best-placed supplier for a particular piece.
How do you handle our confidential information?
Under an NDA signed before the engagement starts. Assessment material stays on our controlled infrastructure, access is limited to the working team, and we return or destroy it at the end on your instruction. We do not use client material as case studies without written permission.
How long does an engagement take?
It scales with the number of functions in scope and how accessible your data and people are. A focused assessment of one department is a matter of weeks. A group-wide strategy across multiple business units takes longer. We give a fixed scope, fee and timeline after a scoping call — no open-ended day rates.
Do you cover AI governance and regulatory compliance?
We cover practical governance: acceptable-use policy, data classification, approval thresholds, audit trails, and the data-residency implications of each platform choice. Where a binding regulatory interpretation is required for your sector or jurisdiction, we work alongside your legal advisers rather than substituting for them — and we will say so rather than overreach.
What if we have already started and it is not working?
That is one of the most productive engagements. A stalled pilot usually has a specific cause — data that was never going to support it, an integration nobody scoped, unclear ownership after launch, or a use case that was chosen because it demonstrated well. We diagnose which, and tell you whether it is recoverable.
Get an Honest Assessment Before You Commit Budget
Free first call — including whether you need a strategy engagement at all
