Measuring whether a drafting tool saved time

If you want to know whether a drafting tool saved your practice time, you have to measure the whole cycle, not the moment the draft appears. The honest unit is hours from instruction to a document you are willing to put your name on, including the hours spent checking facts the tool asserted and the hours spent rebuilding sections you did not trust. Anyone trying to measure time saved on planning drafting without a baseline is comparing a stopwatch against a memory.

Most practices have no baseline, which is the first thing to fix. The second is to accept that the saving may not land where you expected.

What to measure and against what baseline

Pick one document type and hold it constant. A Design and Access Statement for a householder extension and a Planning Statement for a nine-unit scheme are different pieces of work and averaging them tells you nothing.

The baseline is the time your practice currently takes for that document, measured on real jobs rather than estimated. Three to five recent examples is enough to get a range. Record for each one: the fee-earner grade doing the work, the hours booked, the number of draft rounds, and whether the document went out late.

Then run the same document type through the new process for the next three to five jobs and record the same four things. Keep the same fee-earner grade where you can, because a part two architectural assistant producing a first draft that a director then rewrites is a different economics from a director producing it directly.

One caution about the baseline. People remember the worst job, not the median one. A practice that believes a Planning Statement takes a day often finds its own time records show four hours for most and two days for one awkward case.

Drafting hours against checking hours

Split the time into three buckets and record them separately, because they move in opposite directions.

The pattern most practices find is that drafting collapses and verification grows. Whether the total falls depends on how much verification the document needs, which depends on how fact-heavy it is. A Planning Statement reciting local plan policy numbers needs every citation checked. A Design and Access Statement describing a design the practice produced needs far less checking because the practice already knows the facts.

That is the useful insight for scoping work. The documents where the saving is largest are the ones whose content comes from your own project knowledge rather than from external policy that has to be verified.

Rework caused by unverified facts

This is the cost line that destroys an apparent saving, and it is the one practices fail to record because the rework is booked to the original job.

PlanScope's drafts insert a "to be confirmed" flag wherever site-specific facts are missing rather than inventing them, which is a deliberate safeguard and the right behaviour. The risk is the opposite case: a policy reference that looks plausible, reads correctly and is out of date. Policy numbering has changed repeatedly, with the National Planning Policy Framework renumbered in December 2024 and re-issued on 17 August 2026 with policy-coded chapters, so a citation copied from any earlier source needs checking against the current published Framework.

Record rework as its own time entry with a reason code. Three codes are enough: wrong or outdated citation, missing site fact, and tone or structure not fit for the client. After ten jobs the distribution tells you where to put your checking effort. The broader pattern of what AI drafts handle well and handle badly is set out in what drafts get right about structure and wrong about evidence.

Validation failures and their cost

A document that causes an application to be rejected as invalid costs far more than the hours in it. The application has to be remade, the client has to be told, and the programme slips by the time it takes the validation team to look at it again.

Count validation failures before and after, and record the reason each time. If the reason relates to a document the tool produced, that is a direct quality signal. If it relates to a missing document that nobody identified, that is a scoping failure and the tool is not the cause. The common reasons applications come back are catalogued in why planning applications are returned as invalid.

Zero validation failures across a small sample does not prove much, and one failure in twenty does not prove much either. Keep counting over a longer period rather than drawing a conclusion from the first month.

Client perception of quality

Time saved is worthless if the client thinks the document reads as generic. This is harder to measure and worth attempting anyway.

Two usable signals. The number of client comments on a draft, counted per document, which is a rough proxy for how well the first draft met the brief. And whether the client asks who wrote it, which in our field is usually a complaint.

A third signal, if your practice is honest enough to use it: ask the planning officer. Officers rarely volunteer views on document quality and will often answer a direct question about whether a statement addressed the policies they needed addressed. That is the most useful quality measure available and it costs nothing.

Where the saving actually lands

Three places, in descending order of reliability.

Unbilled time comes back first. Narrative documents are frequently written outside the fee, in the evening, by whoever has the file. Reducing that does not increase revenue, and it improves the practice in ways that do not appear in the accounts.

Capacity comes second. A practice that can produce supporting statements faster can take on more applications without adding staff, which shows up as revenue only if there is demand to fill the capacity.

Fee reduction comes third and it is the one to think hardest about. Repricing document production downwards because it now takes less time hands the saving to the client permanently. Some practices do that deliberately to win work. Others hold the fee and bank the margin. That is a commercial decision rather than a technical one, and it should be made consciously rather than by drift.

Where sections are regenerated during drafting, keep the version history straight, because an unrecorded regeneration is a common source of the wrong text reaching the client. The practical approach is covered in version control when report sections are regenerated.

A simple practice log

A spreadsheet with eight columns, filled in at the point of invoicing, gives you everything above.

1. Job reference and document type.
2. Fee-earner grade.
3. Input preparation hours.
4. Drafting hours.
5. Verification hours.
6. Revision hours and reason code.
7. Validation outcome and reason if rejected.
8. Number of client comments on the first draft.

Review it quarterly rather than monthly, because the sample is too small to read sooner. And keep it going after the decision is made, because the measure that matters in year two is whether the verification burden fell as the practice learned where to look.

For documents that need specialist sign-off, the time saving is in the drafting rather than the review, and the reviewer still has to be a qualified person. The professional directory is the route to a reviewer, and the current tiers are set out on PlanScope pricing.

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.