Peer review and the reviewer's scope of liability

Peer review of a planning report can mean three different things, and the difference decides where liability sits. A commentary review produces observations for the author to act on. A verification review confirms that stated methods were followed and that conclusions follow from the data. Adoption means you become the author, with everything that implies. Most disputes about peer review planning report liability start with a reviewer who thought they were doing the first and a client who thought they were buying the third.

So the scope statement is the document that protects you, and it has to be written before the work rather than assembled afterwards.

What a peer review is being asked to do

Ask the client to say, in writing, which of these they want.

A technical commentary, in which you read the report and raise queries, with no statement about its adequacy. Useful in negotiation and worth little as an assurance.

A verification against a named standard, in which you confirm that the method described matches the standard, that the data supports the findings, and that the limitations are properly stated. You are not adopting the survey work and you say so.

A sign-off in which you accept the report as your own, taking responsibility for its content. This normally requires you to have done, or to have directly supervised, enough of the underlying work to stand behind it. For some report types it requires more: the statutory regimes do not permit adoption by someone who was not competent to do the work in the first place.

There is a fourth thing clients sometimes want and should not get, which is a signature on a document you have read quickly for a fixed fee. Decline it. The professional bodies' codes of conduct all require members to act only within their competence, and a signature is a representation of competence whatever the covering email says.

Defining and recording the scope

The scope statement goes at the front of your review, not in an appendix, and it names six things.

1. The documents reviewed, each by full title, revision and date, with a note of the file you were sent.
2. The standard or guidance against which the review was carried out, with edition and year.
3. What the review covered: method, data adequacy, internal consistency, policy citation, conclusions.
4. What it expressly did not cover: calculations not independently re-run, no site visit undertaken, no resurvey, structural or specialist matters outside your discipline, commercial and legal matters.
5. The assumptions you have accepted without checking, and the information you asked for and did not receive.
6. Who may rely on the review, and for what purpose.

That last point is the one reviewers omit. A review addressed to a client and then attached to a planning application is read by the council, the consultees and any objector, and a reliance statement limits who can bring a claim on the basis of it. If the client wants the reviewed document submitted, agree that in advance and price accordingly.

Keep a copy of exactly the file you reviewed. Reports get revised after review, and a reviewer who cannot produce the version they saw is in a weak position when asked about a paragraph that was added later.

Where liability attaches and where it does not

Liability arises in contract, from the terms of your appointment, and in tort, where a duty of care is owed to someone who reasonably relies on your work and suffers loss.

Four practical consequences follow.

Your appointment terms govern most of it, so a written appointment with a defined scope, a liability cap, and a net contribution clause is worth more than any amount of caveating inside the report. Verbal instructions on a fixed fee are the worst position to be in.

Professional indemnity insurance responds to the activities described in your policy. Peer review of another practitioner's work is not always squarely within a policy written for report authorship, so check with your broker rather than assuming, particularly where the report concerns a discipline adjacent to your own.

Limitation periods run from breach rather than from discovery, and the periods differ between contract and tort, and differ again for claims relating to dwellings, where statutory changes have extended the relevant periods considerably. Record retention should follow the longest period that could apply, not the shortest.

And a caveat cannot cure an absence of competence. A reviewer who signs off a fire engineering strategy with a disclaimer about fire engineering has not limited their exposure. They have documented it.

Competent person declarations

Several report types carry a statutory or policy requirement about who may prepare them, and a reviewer needs to know which.

An Environmental Impact Assessment must be prepared by competent experts under the Environmental Impact Assessment Regulations 2017, and the Environmental Statement has to be accompanied by a statement of the relevant expertise. Reviewing such a chapter without that expertise is not a service anyone should offer.

A Biodiversity Net Gain Assessment under the Environment Act 2021 regime expects the statutory metric to be completed by a competent person, with the biodiversity gain plan approved by the local planning authority. Protected species survey and licensing work requires a suitably licensed ecologist under the Wildlife and Countryside Act 1981 and the Conservation of Habitats and Species Regulations 2017. Land contamination reporting is expected to follow BS 10175 and to be prepared by suitably qualified and experienced persons. A Fire Statement sits within the Building Safety Act 2022 framework and Planning Gateway One and needs fire engineering competence. In each of these the AI-assisted or junior-authored draft is a starting point for that professional rather than a substitute for them.

Where a report type is labelled as professional recommended rather than required, such as an arboricultural impact assessment to BS 5837:2012 or a noise assessment to BS 4142:2014+A1:2019, councils usually expect a suitably qualified author and professional review is strongly advised.

State your own competence explicitly: qualifications, chartered or licensed status, membership grade, and the number of comparable reports you have reviewed or authored. It reads as unusual the first time you do it and it is the single most useful paragraph in the document.

Reviewing an AI-assisted draft

The work is different from reviewing a junior's draft, because the failure modes are different.

Check provenance first. Every number in the report should trace to a named source, a survey with a date, or a calculation you can follow. Where a draft cites policy, verify the citation against the current published Framework edition rather than accepting it, because policy has been renumbered more than once and drafting tools trained on earlier editions reproduce old numbering.

Then look for what is missing rather than what is wrong. Well built drafting tools flag gaps instead of filling them, and PlanScope inserts a "to be confirmed" flag wherever site-specific facts are absent. Those markers are the useful part of the document. A draft that reads as complete with no gaps at all, on a site nobody has visited, deserves more suspicion than one with twenty flags in it.

Check that no survey is implied that was not carried out. The sentence to hunt for is the one that describes a site condition in the present tense without a survey date attached.

Then read the conclusions against the data, which is the ordinary discipline of any review, and the same approach that applies to a mechanical schedule such as the condition grading in a structural survey defect schedule.

Wording a sign-off that means something

A declaration should say who you are, what you reviewed with revision and date, what standard you applied, what you did and did not do, what you concluded, and on what basis you are competent to say it. Sign it and date it.

Avoid three words. "Approved", because it implies an authority you do not have. "Compliant", unless you name the thing complied with and the version of it. And "satisfactory", which means nothing without a criterion.

A declaration that says the report has been reviewed against a named standard, that the method and data support the stated conclusions, that identified matters have been addressed in the revision listed, and that specified elements were outside the scope, is a document you can stand behind three years later.

Records to retain after sign-off

Keep the reviewed file as received, your mark-ups, the correspondence in which scope was agreed, the information requests and the responses, the final declaration, and evidence of your competence at the date of signing, including membership and continuing professional development records.

Retain them for as long as a claim could be brought, which for work relating to dwellings may be considerably longer than the six years people assume. Storage costs almost nothing set against reconstructing a review from memory.

PlanScope's professional directory is the route through which drafts are sent for review and sign-off, with request statuses running from sent through accepted, in review and changes requested to signed off. PlanScope provides AI-assisted draft reports for planning applications. Reports should be reviewed before submission. We do not guarantee planning approval. For specialist technical reports, qualified professional input may be required.

This is general guidance, not legal advice; check the position with a qualified professional or your own insurer before relying on it.