Why estimators lose days re-keying BOQs, and what to do about it

Ask a tender team where the first days of a bid go and the answer is rarely pricing. It is copying: lifting items, descriptions, units and quantities out of the client’s BOQ, which arrives as a locked PDF, a scanned schedule or an Excel file laid out differently from every other one, and re-keying them into the company’s own pricing sheet.

Why re-keying is more than a time problem

The obvious cost is hours. The less obvious cost is what those hours crowd out. In a widely cited 2018 survey of US construction professionals by FMI and PlanGrid (now Autodesk), respondents reported spending about 35% of their time, over 14 hours a week, on non-productive work such as looking for project information and dealing with mistakes and rework. Re-keying a BOQ is that kind of work: necessary, but it adds nothing the client is paying for.

Experienced estimators point at what gets squeezed. In his video on the five worst estimating mistakes, estimating coach Tim Fairley highlights ignoring the contract conditions in favour of the drawings, and missing small items of scope that add up. Both happen when the team spends its time typing and has little left to read the documents properly.

Where the errors come in

• Units changed in transit (m² vs m, nr vs item, lump sums split or merged)
• Item references lost, so the priced BOQ no longer maps back to the client’s numbering
• Descriptions shortened, dropping the clause that defines what is included
• Addenda and revised BOQs applied to one sheet but not another
• Provisional sums and prime cost items priced as if they were measured work

What practitioners recommend

The advice from estimators and estimating guides is consistent, and most of it comes before any software decision.

• Fix one internal BOQ structure: your columns, your trade or section codes, your units list. Every incoming BOQ is mapped into it, not the other way round.
• Keep the client’s item reference on every line so the priced BOQ can always be traced back and returned in their format.
• Separate quantities from rates. Quantities come from the tender; rates come from your library and supplier quotes. Mixing them in one pass is where mistakes hide.
• Log every addendum against the lines it changes, with a date.
• Run a check before pricing starts: line count, section totals and units compared against the source document.

Where AI-assisted extraction fits

Reading a tender BOQ and turning it into structured rows is one of the more realistic jobs for AI today, because the output can be checked line by line against the source. It is not a reason to skip review. Independent testing of AI takeoff tools, such as the Utica University study we covered in an earlier guide, found good results on simple counts but weaker accuracy on harder quantities, with estimator oversight still essential. The same principle applies to extraction: treat the output as a first draft, and keep the source reference on every line so a person can confirm it quickly.

A sensible pattern is extraction into your standard structure, automatic flags for anything unusual (unit mismatches, missing quantities, items that look like provisional sums), then an estimator reviewing the flagged lines before pricing begins.

How we approach it

We build the extraction around your own BOQ template and units, link every line to its source page, and leave approval of the structured BOQ with your estimator before any rate is applied.