Knowledge / AEC-native data

Bills of quantities

The measured, itemised list of work and materials priced by tenderers, structured to NRM2 or a bespoke measurement rule set.

The bill of quantities is the measured list of what tenderers are pricing: every item of work and material, measured by the quantity surveyor to NRM2 or a bespoke rule set, organised under element and work section headings, with references to the clauses and drawings it was measured from. It arrives as a PDF or spreadsheet of several hundred pages with the tender pack, is corrected by addenda, then fixed as the contract bill for valuation after award. A bill of quantities AI has to answer what is and is not measured, which sums are provisional, and where an item disagrees with its drawing. Structured knowledge makes the whole bill something the commercial team can question item by item, with each answer traceable to its page, heading and rule.

What teams need to ask of it

  • Is this item of work measured, deemed included in another item, or covered by a provisional sum?

  • Which provisional and prime cost sums are in the bill, what do they cover, and which drawings and clauses do they refer to?

  • Which items reference a specification clause or a drawing that has since been revised?

  • Where do the measured quantities look wrong against the drawings they were taken from?

  • What changed between the tender bill and the contract bill, and which addendum changed it?

  • Which items carry a measurement note (deemed included, extra over, method of measurement) that changes what the rate must cover?

  • Has the same work been measured twice across packages, or fallen between them?

Why generic AI gets it wrong

A bill item means nothing on its own. "Ditto, but 150 mm thick" takes its meaning from the item above; the item above takes its meaning from the sub-element heading, which takes its meaning from the element and work section headings above that. Flatten the bill to text and the hierarchy goes, so a tool reads a line about excavation without knowing which foundation it belongs to, or reads a rate item without the preamble that states what the rate is deemed to include. Measurement rules carry meaning a generic reader has never learned. Under NRM2, certain work is deemed included in other items and is not measured separately, and a model that has not read the preambles will report it as missing.

The bill also exists in at least two states. The tender bill goes out with the ITT, is corrected by addenda and tender queries, and is then re-issued as the contract bill with the tenderer's rates. A tool given one version quotes the quantity from that version, without knowing that addendum three reduced it or that a post-tender clarification moved the item into a provisional sum. The answer comes back with an item reference and a quantity. It looks like the bill. It is the wrong bill, and a valuation built on it is wrong in a way nobody notices until final account.

One question, answered

QUESTION

ANSWER

SOURCES

See it on your data.

See it on your data.

We onboard a focused slice of your project data, configure the ingestion paths, extract a structured context graph, and build a workflow against the resulting structured knowledge layer.

Related to

The bill is measured from the construction drawings and construction specifications and issued as part of the tender packs and ITT documents, though on design and build projects its place is usually taken by a contract sum analysis or pricing schedule read with the employer's requirements. Tender document review uses it to test what the pack is asking tenderers to price, and the variation and change impact review returns to it after award, when a change has to be valued against the contract rates.

Frequently asked questions

Frequently asked questions

Does it work with a bespoke measurement rule set, or only NRM2?

Which formats does it read?

Can we test it on one of our own bills?

Related pages