Home / Blog

Estimating and quoting records: the data AI developers find hardest to get

Construction drawings, tape measure, calculator and laptop

Most of what a business publishes shows the finished result. What it almost never publishes is how it got there: how a job was measured, how it was priced, what was assumed and what changed. That’s exactly what estimating and quoting records capture.

Why it’s rare

Pricing is commercially sensitive, so it rarely appears anywhere public. That makes real estimating records, with their working, scarce for anyone building tools to help with quoting, planning or cost control.

Records that tell the story

  • Enquiries and the drawings or specifications received
  • Take-offs and measurement sheets
  • Quote build-ups, rates and assumptions
  • Revisions, clarifications and value-engineering notes
  • Whether the quote was won or lost, and what the job eventually cost

The link to outcomes

A quote on its own shows a price. A quote linked to the final account, variations and job notes shows how accurate the estimate was and why. If your estimating software, spreadsheets and job files can be kept together, they’re worth much more as a set.

What stays out

Customer names and personal details come out. You can leave out particular customers, frameworks or years. Supplier terms that are confidential can be excluded too.

Old quotes still count

Records from older years can be just as useful as recent ones, because they show how methods and costs changed. Please don’t clear them out just because the jobs are finished.

See also what AI labs actually want from your business records.

Questions about a particular archive? Email [email protected] or tell us what you’ve got.

Got records you think might be worth something? A short description is enough for a free first look.