The Prompt and the Judgement: Commissioning Advice When the Board Has Already Consulted a Machine
Photo: Jeremy Segrott from Cardiff, Wales, UK, CC BY 2.0, via Wikimedia Commons
By 2026, it is unremarkable for a UK board pack to arrive with an AI-generated market summary attached. A finance director may have used a generative tool to draft a strategy note. A chief executive may have asked for a competitor scan before the first meeting with a consultant. The machine has already been in the room, even if it was not invited.
This changes the terms of a consulting engagement before it begins. Clients are less likely to pay for the production of information that they can generate themselves. They are more likely to expect synthesis, challenge, and a clear route through competing options. Consultants who present polished prose as the core deliverable will find that the client has a faster, cheaper version of the same output.
From information to judgement
The useful distinction is between generative fluency and situated judgement. Generative tools are good at producing plausible language, summarising public material, and offering a wide range of options. They are less good at knowing which option will survive a boardroom argument, which stakeholder will object, or which operational constraint will delay implementation.
For UK SMEs and mid-market boards, the risk is not that AI produces nothing useful. The risk is that it produces something plausible enough to create false confidence. A generated strategy can look comprehensive while missing the history of a family business, the politics of a private equity-backed board, or the regulatory nuance that sits outside the public record.
What clients should commission
The commissioning conversation should start with the decision to be made, not the document to be produced. A useful brief will ask the consultant to identify the trade-offs, test the assumptions, and recommend a course of action that someone can be accountable for.
That leads to a different set of questions in the proposal. Who will sign the recommendation? What evidence sits behind it? Which assumptions have been tested with the people who will have to implement it? What is the cost of being wrong? How will the board know whether the advice is working?
AI can support this work. It can speed up analysis, generate alternative scenarios, and challenge a first draft. It cannot carry the accountability for a recommendation, read the room during a difficult board discussion, or negotiate between factions. Those are the parts of consulting that remain stubbornly human.
A practical test for buyers
When a proposal arrives, the buyer can ask three questions that separate generated content from genuine judgement.
First, what is the evidence trail? Ask the consultant to show how they moved from data to conclusion, including where they used AI and where they overrode it. A transparent method is more valuable than a confident answer.
Second, who is accountable? The proposal should name the person who will stand behind the recommendation. If the work is a team effort, the client should know which judgement sits with whom.
Third, what cannot be known from the public record? This is the question that tests whether the consultant has actually engaged with the client's context. The answer will reveal whether they understand the stakeholder map, the internal constraints, and the political tensions that no generated summary will capture.
The human in the room
The consultant's value in 2026 is not in producing more text than the client's AI tool. It is in knowing which text matters, which trade-off is acceptable, and which recommendation will still make sense when the boardroom temperature rises. UK clients should commission that judgement explicitly. Anything less is a document, not advice.
