The Assurance Memo: Stress-Testing a Consultant’s Recommendation Without Duplicating the Work

A second opinion is often imagined as a repeat exercise: another adviser, another discovery phase, another fee. That is rarely necessary. When a recommendation is significant but the underlying work has already been done, the client’s need is narrower. It is an independent check on the logic, evidence and conditions behind the advice. In 2026, with consultancy budgets under scrutiny and boards asking harder questions, the practical question is not whether to re-run the project. It is how to test the recommendation without dismantling the relationship that produced it.

When a second opinion earns its cost

A second opinion is worth paying for when the cost of being wrong is materially higher than the cost of checking. That usually means the decision is expensive to reverse, the advice rests on forecasts or assumptions, or the adviser has a commercial interest in the next phase of work. It also applies where a board, lender or regulator needs evidence that the recommendation has been challenged before approval.

There are quieter triggers. The internal team may lack the specialist knowledge to interrogate a technical claim. The recommendation may conflict with operational experience but no one can articulate why. Or the adviser may be the only person who has seen the full evidence base. In each case, a narrow independent review can supply the missing challenge.

A second opinion is less useful when the issue is a straightforward factual error that the original adviser can correct, or when the decision is low-value and reversible. It is also a poor substitute for a clear brief. If the original scope was vague, the first task is to fix the scope, not to commission another adviser to interpret the ambiguity.

Ask the original adviser to expose the evidence

A useful challenge begins with a request for the evidence pack. This is not an accusation. It is a structured request for the material that any competent adviser should be able to produce. The pack should include the assumptions register, the data sources and their limitations, the options considered and discarded, the sensitivity analysis, and the conditions attached to the recommendation.

It should also disclose conflicts, dependencies and dissent. If the adviser’s firm is likely to deliver the implementation, that should be stated. If a key conclusion depends on a single customer, supplier or regulatory judgement, that should be visible. If a member of the project team disagreed, the client should know the substance of that disagreement rather than simply its existence.

For modelling work, ask for the model and a short note on how it behaves under stress. For strategic advice, ask for the decision log: what was considered, what was rejected, and why. The aim is not to demand a second report. It is to make the original reasoning inspectable.

Frame the challenge before it becomes a confrontation

How the request is framed matters. A challenge memo should separate clarification from justification. Some questions simply ask the adviser to explain a term or source. Others ask them to defend a judgement. A smaller number ask them to test an alternative scenario. Mixing these together invites defensiveness.

A practical approach is to send the questions to the original adviser first, with a reasonable response time and a clear format. This gives them the chance to correct misunderstandings and to supply missing evidence. It also creates a record. If the client later seeks an independent view, the reviewer can see what was asked and what was answered.

Confidentiality should be agreed at the outset. The client should be clear about who will see the challenge memo, who will see the independent review, and whether the original adviser will have a right of reply. Data protection obligations should be respected, particularly where personal or commercially sensitive information is shared.

Brief the reviewer to answer questions, not to rebuild the project

The independent reviewer should be given a narrow brief. The instruction is not to review the project in full. It is to answer a small set of questions, usually between four and six, with reference to the evidence provided. Those questions might include whether the recommendation follows from the analysis, whether the assumptions are reasonable, whether material risks have been omitted, and whether the proposed conditions are enforceable.

The reviewer should be asked to state what would change their view. That single request separates a genuine assurance exercise from a vague expression of comfort. It also helps the client see the limits of the opinion.

A fixed fee and a short timetable are appropriate. Many assurance memos can be completed in one to three days if the evidence pack is ready and the questions are precise. The output should be concise: a short written opinion, a summary of evidence reviewed, any qualifications, and a clear statement of what the reviewer did not examine. The client should resist the temptation to expand the brief once the review begins.

Use the findings to strengthen the decision

The value of a second opinion lies in what the client does with it. There are four sensible responses. The client can accept the recommendation with conditions. They can qualify approval pending further evidence. They can pause the decision while a specific risk is tested. Or they can reject the recommendation and record why. Each response should be documented, with the reasoning and any dissent noted.

The exercise should not become a way of avoiding accountability. A second opinion informs the decision; it does not make it. The client still owns the judgement. Nor should it be used to embarrass the original adviser. A well-run assurance process usually improves the quality of the original advice, because it forces the reasoning into the open while there is still time to act on it.

Common pitfalls to avoid

There are recurring mistakes. One is appointing a reviewer from the same methodological school as the original adviser, which can produce agreement rather than challenge. Another is briefing the reviewer so broadly that the exercise becomes a duplicate project. A third is framing the request adversarially, which may secure a defensible memo but damage the working relationship.

A fourth is failing to keep a record. If the board later asks why the recommendation was approved, a file containing the challenge questions, the adviser’s response and the reviewer’s opinion is more useful than a recollection that someone once looked at it. A fifth is treating the review as a substitute for management judgement. It is not.

A proportionate discipline

The assurance memo is a proportionate discipline for important decisions. It respects the work already done, tests the recommendation rather than the relationship, and gives the client a documented basis for approval. For UK organisations facing complex advice and limited internal challenge, that is often a better use of money than commissioning a second project. The aim is not to duplicate the consultancy. It is to make the original advice stand up to scrutiny before the client commits.