Reviewing an AI drafted report as the named professional

When you sign off a draft you did not write, you become the author for professional purposes. Reviewing an AI drafted planning report is therefore not proofreading. You are adopting every citation, every factual assertion about the site and every methodological claim as your own, and your institute's code of conduct, your professional indemnity insurer and the local planning authority will all treat you accordingly.

That is manageable, and in some ways easier than reviewing a junior colleague's work, because a well constructed draft flags what it does not know. The risk is the draft that does not.

What you are taking responsibility for

Be specific with yourself about the scope. Signing a report means you are stating that the methodology is appropriate for the question, that the standards cited are the right ones and are current, that the site-specific facts are correct, and that the conclusions follow from the evidence.

It does not mean you carried out the survey. Where the report relies on fieldwork you did not do, say so in the declaration and name who did it and when. A review that silently adopts someone else's survey data is the position most likely to go wrong later.

For report types where a licensed or competent professional is legally required, including biodiversity net gain assessments, protected species surveys, land contamination work, environmental statements and fire statements, the draft is a starting point for your work rather than a document you can simply approve. The professional directory sets out the review route, and the status of a review request moves through a defined sequence from sent and accepted, through in review and changes requested, to signed off, rejected or expired. Use changes requested rather than signing with reservations.

Checking citations and editions

This is where most problems sit, and it is quick to check.

Policy references age badly. The National Planning Policy Framework was renumbered in December 2024 and re-issued on 17 August 2026 with policy-coded chapters in place of paragraph numbers, so a draft citing a paragraph number may be quoting an edition two versions out of date. Verify every policy reference against the current published Framework and state the edition in the text.

Then work through:

1. Every statutory reference, checking the full title, year and section against the current consolidated text, since amendments are frequent.
2. Every British Standard and professional guidance document, checking the edition and year, because the methodology may have changed between editions.
3. Every local plan policy, checking the policy number against the adopted plan rather than a superseded draft.
4. Every numerical threshold, since thresholds move and a figure that was right in a previous edition may not be.
5. Anything presented as a statistic, which should either have a named source or be removed.

A draft that cites a paragraph number without an edition is telling you it has not been checked. Treat that as a flag rather than an irritation.

Verifying site specific facts

The site facts are the part a drafting system cannot know, and the part a reader will test. Check each against primary evidence rather than against the draft's internal consistency.

The address, the red line area, the existing use and its lawful basis. The flood zone, taken from the current official mapping rather than a prior report. Designation status for heritage, ecology and landscape, taken from the national registers. Levels and dimensions, taken from the topographic survey. The planning history, taken from the authority's own register.

Where a fact in the draft cannot be traced to a source document, either source it or remove the sentence. The temptation to leave a plausible statement in place because it is probably right is the single habit that causes reviewers trouble.

Spotting a gap the draft has filled

A properly constructed draft inserts an explicit flag, such as a bracketed note that a fact is to be confirmed, wherever site data is missing, and that behaviour is a safeguard rather than a defect. Your job is to find the places where that did not happen.

The patterns to look for:

• A confident sentence about the site that no appendix supports.
• A conclusion stated more firmly than the evidence allows, particularly in the balancing or impact sections.
• A methodology described in general terms where the standard requires a specific procedure.
• Mitigation proposed without a design, a location or a responsible party.
• Any number in the document that is not traceable to a survey, a drawing, a published schedule or a cited source.

Where the draft has substituted generic policy language for a site-specific judgement, that is the sentence to rewrite. Generic text reads smoothly, which is exactly why it passes unnoticed.

Scope of review against scope of authorship

Decide which of the two you are being asked for, and write it down before you start.

A review means the client keeps authorship and you give a written opinion on the document, listing the changes required. Your name appears as reviewer, not as author. Your liability is for the opinion.

Authorship means you take the document over, make the changes yourself and sign it as your work. Your liability is for the whole report.

The middle ground, where you make extensive changes but describe yourself only as a reviewer, is the worst place to be. Pick one and reflect it in the engagement letter, the declaration in the report and your insurer's understanding of the work you do.

The same question arises wherever authorship is split, and the reasoning is set out in the notes on construction management plan author accountability and drainage engineer competence.

Recording the review

Keep a review record that would make sense to a third party two years later. At minimum: the version reviewed with its date and a copy retained, the checks carried out, the issues raised, the client's response, the changes made and the version signed.

Mark up the document rather than sending a summary email. A tracked set of changes is evidence of what you did. A message saying the report looks fine is evidence of nothing.

Where your opinion differs from another consultant's on the same question, narrowing the dispute in writing is more productive than two reports in opposition, and the approach is covered in the note on statements of common ground and expert conferencing.

Declining to sign, and the fee

Sometimes the right answer is no. The usual reasons are that the report depends on survey work that has not been done, that the survey season has been missed, that the methodology is wrong for the question, or that the client will not accept a conclusion you consider correct.

Say it plainly and in writing, give the reason, and say what would change your view. A clear refusal with a route forward is better for the client than a qualified signature, and far better for you.

On fees, charge for review by time rather than as a percentage of the drafting fee, because the effort depends on the state of the draft rather than its length. Price the verification work separately from any fieldwork, and be explicit that a second round of review is chargeable. A draft arriving with its citations already checked takes a fraction of the time of one that does not, and your fee basis should reflect that honestly.

See the full index of report types for the governing standard and professional requirement attached to each, then draft and send for review on PlanScope.

This is general guidance, not legal advice; check the position with your local planning authority, your professional institute or your insurer before relying on it. 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.