Signing off an AI-assisted draft: what you take on

The moment you sign, the report is yours. Not the drafting tool's, not the applicant's, and not the junior who assembled the appendices. A professional review of an AI-drafted planning report is an act of authorship with a shorter run-up, and the scoping conversation you have before accepting the work decides whether that is comfortable or dangerous.

That framing matters because the volume of AI-assisted drafts reaching reviewers is rising, and the commercial pressure is to treat review as a fast read. It is not.

What signing actually means

Your signature tells a case officer that a person with the relevant competence has checked the content and stands behind it. Institute codes of conduct across the RTPI, CIEEM, the Institute of Acoustics, CIHT and IStructE are built on the same two ideas: act only within your competence, and be honest about the basis of your conclusions.

Neither idea has an exception for text you did not type.

For a handful of report types the position is statutory rather than professional. An Environmental Statement must be prepared by competent experts under the Town and Country Planning (Environmental Impact Assessment) Regulations 2017, and the statement is accompanied by evidence of that expertise. Protected species survey work and any mitigation licence sit with a suitably licensed ecologist under the Wildlife and Countryside Act 1981 and the Conservation of Habitats and Species Regulations 2017. Biodiversity net gain under the Environment Act 2021 expects the statutory metric to be completed by a competent person, with the biodiversity gain plan approved by the local planning authority. In those cases a review does not create competence that was absent when the draft was made. The work has to be done by the right person.

Scoping the review before you accept it

Ask for the whole picture before quoting. A draft Environmental Impact Assessment chapter with no baseline monitoring behind it is a different job from a draft with a full survey dataset attached.

The questions worth asking are short.

1. What survey or monitoring data exists, when was it collected, and by whom?
2. Which standard or guidance does the assessment claim to follow, and is the correct edition being used?
3. Has a site visit happened, and can you get access to do your own?
4. What is the submission date, and does it leave room for a survey season you may need?
5. Who else has touched the document, and what will the final version look like when it is submitted?

Then set a stop point. A short scoping fee, followed by a written note saying either that the draft is reviewable or that the work needs to restart, is far easier than discovering the problem three days before validation.

Checking facts, figures and citations

Work through a draft in a fixed order rather than reading it front to back. Policy citations first, because they are quick to check and a wrong edition undermines everything downstream. The National Planning Policy Framework has been re-issued more than once in recent years and the current edition is organised by policy code rather than by the paragraph numbers many older documents quote, so any citation copied from an earlier draft needs verifying against the published Framework.

Next, numbers. Every figure should trace to a source you can open: a survey sheet, a gauged record, a published dataset, a drawing. Numbers that cannot be traced come out, and the sentence is rewritten to describe the effect without a figure.

Then method. A draft can describe a method perfectly and still be describing the wrong one for the site, and only a specialist reader catches that. Check the assessment approach against the standard the report claims to follow, and check that the standard is the current version.

Finally, the conclusions. Read them cold, as if you had not read the rest, and ask whether the evidence in the document actually supports them. Drafts tend to be more confident than their evidence base.

Site-specific gaps a draft cannot fill

Well-built drafting tools flag missing site facts rather than inventing them, which is the safeguard you want and also the thing that generates your work list. A draft that leaves a marker where the winter water table should be is telling you the truth.

The gaps that recur are the same across disciplines. No site visit and therefore no record of what is actually there. No seasonal survey, which for bats means emergence work between May and September and for great crested newts means environmental DNA sampling between 15 April and 30 June. No baseline measurement where a standard requires one. No consultation response from the statutory body. No calculation run in the correct tool, which for a Biodiversity Net Gain Assessment means the statutory metric rather than a description of it.

Each gap has three lawful endings: fill it, commission it, or state it as a limitation. There is no fourth.

Recording what you did and did not verify

Keep a review record separate from the report itself. It does not need to be long, but it needs to name the version you reviewed, the date, the documents you were given, the checks you carried out, the items you could not verify and the changes you required before signing.

Two things make that record useful later. First, it fixes the version. Drafts get edited after review, and a record that identifies the exact file you approved is the difference between a straightforward conversation and an argument. Second, it captures your limitations in your own words. A Protected Species Survey signed on the basis of a single visit outside the optimal season needs that fact on the face of the report, not buried in correspondence.

Where a report is submitted with residual uncertainty, say so plainly in the document. Councils deal with uncertainty every day. They deal badly with discovering it after a decision.

A short standing form helps here. List the document title and version, the date of your review, the material you were supplied, the site visit date or its absence, the standards you checked the method against, the specific figures you traced to source, the items you were unable to verify, and the changes you required. Store it with the report rather than in an email thread. Reviewers who keep this record find that the awkward questions arrive years later, long after the correspondence has been archived, and by then the record is the only thing that remembers what you actually saw.

One more habit worth adopting: review the version that will be submitted, not an earlier one. Ask for the final assembled document including appendices and drawing schedule, and sign that. Where the applicant wants to add material afterwards, treat it as a new review rather than an amendment to yours.

Insurance and fees

Check the professional indemnity position before the first review, not after the first claim. Policies differ on work produced by others and later adopted, and some ask about the proportion of your fee income that comes from checking third-party documents. A short conversation with your broker, recorded in writing, costs nothing and settles the question.

On fees, the honest structure is a scoping fee, then a review fee, then a separate quote for anything the review shows is missing. Reviewers who quote a flat rate for signing off drafts end up subsidising the worst ones. And a stated exclusion in your terms, saying that the fee covers review of the draft as supplied and not the collection of new survey data, prevents most of the awkward conversations.

Where you decide the draft cannot be brought to a signable state, say that in writing and stop. Declining is part of the job, and it is much easier to decline before your name is attached than afterwards.

If you review planning documents professionally, PlanScope's professional review route is where sign-off requests are handled, and the full list of report types shows which ones are labelled as requiring a competent professional rather than merely recommending one.

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.