The Specialist Premium: What UK Boards Should Still Pay For as Generalist Advice Becomes Automated

Artificial intelligence has made generalist advice cheap. That is not a controversial claim in 2026. Board papers, market summaries, options analyses and regulatory overviews can now be produced at speed and at low cost. The risk is not that these outputs are useless. The risk is that fluency is mistaken for insight, and that boards pay specialist fees for work that is essentially well-formatted generalism.

The more important consequence is subtler. As broad consultancy output becomes commoditised, genuine specialist judgment should become more valuable, not less. The premium is no longer for information. It is for interpretation, accountability and evidence of delivery in a specific UK context.

The new commodity is confidence

AI tools are good at producing confident prose. They are less good at knowing which questions a regulator will actually ask, which trade-off a board can defend to its stakeholders, or which seemingly minor contractual point will determine whether a transaction completes. Those judgments depend on pattern recognition earned through delivery, not on the breadth of a model.

Boards should therefore stop treating a polished document as evidence of expertise. The relevant question is not whether an adviser can summarise the landscape. It is whether the adviser can tell the board what matters, what does not, what they would do and what they would refuse to do.

Demand proprietary sector insight

Specialist advisers should bring something the board cannot easily assemble. That may be sector-specific benchmarking, named experience of regulatory engagement, live market intelligence, transaction precedents or a track record of implementation in comparable organisations. Proprietary insight is not the same as a licensed database. It is the practical knowledge that comes from having operated in the sector and having been accountable for outcomes.

In procurement terms, boards should ask for evidence rather than assertion. Which named individuals will do the work? What comparable assignments have they delivered? What did they recommend, and what happened next? Anonymised case examples are acceptable, but generic capability statements are not. If the insight could have been generated by a competent generalist with an AI assistant, it is not specialist insight.

Require accountable recommendations

One of the clearest distinctions between specialist advice and automated generalism is accountability. An AI tool cannot sign a recommendation, carry professional liability or sit in front of a board and defend a judgment under challenge. A specialist adviser can and should.

Boards should require a clear recommendation where the brief calls for one. An options paper that lists possibilities without a view is often an abdication of judgment. The adviser should state the recommendation, the assumptions behind it, the principal risks, the alternatives considered and the consequences of being wrong. The firm should identify the accountable partner or lead and confirm liability, insurance and escalation arrangements.

This is not a technicality. It changes the nature of the purchase. The board is buying judgment that can be examined, challenged and owned.

Regulatory nuance is not generic compliance

UK regulatory advice is particularly vulnerable to AI-assisted generalism because the rules are easy to summarise and harder to apply. FCA Consumer Duty, PRA expectations, UK GDPR and ICO practice, CMA competition analysis, the Procurement Act 2023, national security investment screening and employment-related obligations all depend on context. The same rule can produce different consequences depending on sector, customer base, governance history and regulator relationship.

Specialist advisers should be able to explain how the regulator behaves, not merely what the rule says. They should offer a regulatory route map, identify likely areas of challenge and explain what evidence the organisation should retain. Boards should be wary of advisers who present compliance as a checklist. The value lies in knowing where the checklist is incomplete.

Look for evidence of delivery

Advice is not the same as implementation. Boards should be clear about which they are buying. If they are buying advice, the adviser should still be able to show that previous advice was actionable and was acted upon. If they are buying delivery, the evidence standard is higher: named delivery leads, references, post-implementation reviews and a willingness to be measured against agreed outcomes.

A specialist adviser should not be afraid of a post-engagement review. Nor should they resist knowledge transfer. A good adviser leaves the organisation more capable than they found it. A weaker one creates dependency.

Telling real expertise from AI-assisted generalism

Real expertise often looks narrower than generalist confidence. Specialists ask unusual questions. They volunteer bad news early. They identify trade-offs that the board had not considered. They can explain what they would not do, and why. They bring their own data, precedents and sector relationships. They are comfortable saying that the answer depends on facts the board has not yet supplied.

Red flags include generic frameworks applied without adjustment, unnamed delivery teams, implausibly fast turnaround on complex issues, citations without provenance and reluctance to accept accountability. A adviser who cannot distinguish between a sector-specific judgment and a plausible summary is unlikely to provide the assurance a board needs.

Procurement implications

The procurement response should reflect the nature of the purchase. Boards should define the outcome, require a method, name the personnel, address conflicts, confirm data protection and AI use, secure knowledge transfer and set clear liability. Buying hours rewards activity. Buying judgment rewards outcomes.

Competitive tension is useful, but a reverse auction is a poor way to procure high-risk specialist advice. The cheapest plausible proposal may simply be the most automated one. The board should test whether the adviser can defend the recommendation under scrutiny, not merely whether they can produce it quickly.

The premium that remains

AI has raised the floor of generalist advice. It has not raised the ceiling of specialist judgment. UK boards should pay for the latter: sector-specific insight, accountable recommendations, regulatory nuance and demonstrable delivery. The test is simple. If the advice cannot be challenged, owned and evidenced by a named specialist, the board may be buying output rather than assurance.