The board wants an AI plan
Owners or the board ask what the company is doing about AI, and nobody owns the answer. You need a position you can defend, not a vendor’s roadmap.
AI strategy discovery · five days · fixed price
Your AI plan, decided in five days.
For leadership teams being asked “what is our AI plan?”. Five working days of interviews and workflow review, spread over about two weeks, end in a decision you can act on: which use case comes first, what not to do, who owns it and what it will cost to run. Fixed price, from £3,000.
Led personally by Radko Diev — former COO and product owner of a regulated platform. No vendor or tool is chosen in advance.
AI plan: the first 90 days
- Do first
- Supplier-invoice matchingOwner: finance lead · baseline: today’s cycle time
- Do next
- Customer-service triageOwner: head of service · once the first is live
- Not now
- Contract-review assistantReason: the contract register is not digitised yet
Signed by the sponsor · reviewed on day 90
Who it is for
Three situations it was built for.
Too many ideas, no basis to choose
Suggestions from vendors, staff and competitors pile up. Each one is scored on value, effort, risk and running cost, so the first move can be defended.
Pilots that never went live
A pilot ran and stalled. The same five days find what actually blocks production — data, integration, ownership, running cost or the model — and decide: scale, fix or stop.
How the five days run
Five working days, spread over about two weeks.
Interviews need diaries, so the work is spread out. The sponsor gives about four hours — the kick-off and the decision workshop — and each person interviewed about an hour, up to two where a workflow is walked through.
- Day 1
Kick-off
The questions that must be answered, confirmation of the agreed scope — one function, two or three, or the whole organisation — and access to the documents and systems that matter.
- Days 2–3
Interviews and walk-throughs
Three to six conversations with the people who run the work: where it waits, gets re-keyed or is checked too late; what data exists and who owns it; what has been tried.
- Day 4
The scored portfolio
Every candidate scored for value, feasibility on the data you already hold, risk and the oversight it needs — with a named owner, the baseline it must beat and the cost to run.
- Day 5
The decision workshop
Two hours with leadership: choose the first one or two, agree what not to do, and walk through the 90-day plan.
What you receive
Four documents your team keeps.
Written and editable, in English or Bulgarian, plus a verdict on each pilot where pilots are in scope. No slide deck unless you ask for one.
Decision page
One page leadership can sign: what to do first, what not to do, and who owns it.
Use-case portfolio
Every candidate scored for value, effort, risk and running cost against today’s baseline.
90-day plan
Owners, dependencies and the date each decision falls due.
Readiness view
Data, systems, controls and people — for the use cases chosen, not a generic maturity score.
Illustrations of the formats, not client material.
What it costs
Fixed price, agreed before work starts.
- One function or department
- £3,000
- Two or three functions
- £4,000
- The whole organisation, up to about 300 people
- £5,000
- Larger organisations
- Quoted after a call
Agreed in writing before work starts; prices exclude VAT. Delivered remotely with expenses included; on site by agreement. Clients in the EU are quoted in euros. A working session in the previous 30 days is credited in full, and the discovery fee is credited against a Sprint commissioned within 30 days.
Not ready for five days? A one-day working session, from £1,500, takes a single process instead. See the working session
FAQ
Common questions.
How much of our time does it take?
About four hours from the sponsor — the kick-off and the decision workshop — and about an hour from each person interviewed, up to two where a workflow is walked through. Sofpact does the rest.
What do you need from us to start?
A sponsor who can make the decision, three to six people who run the work, and read access to the documents or systems that matter. The discovery needs no extracts of customer or staff records, and nothing is copied out of your systems without your agreement.
What if the answer is “not yet”?
Then that is the result, and the decision page says why — for example, data that is not ready, or a process that has to change first. Stopping on evidence costs less than a pilot that stalls.
Do you favour particular vendors?
No. No platform, model or automation tool is chosen in advance, and recommendations follow your data, risk and running cost. Any vendor relationship that could touch your work is declared in the scope.
Who does the work?
Radko Diev, personally — every interview, the scoring and the workshop. If the plan leads to a build, Radko builds the first version in a Sprint; any engineers it needs are named and agreed with you in the scope.
Is this legal or compliance advice?
No. Each use case gets an operational risk tier and the human oversight it needs; your counsel and data protection officer own the legal interpretation.
Further reading How to build an AI strategy that survives contact with operations
A useful first conversation