The situation
The business is a fit-out and refurbishment contractor in the north of England with around 30 staff on the books and a steady reputation built over fifteen years. Most of the work comes in through repeat clients, agents, and word of mouth, and the tender pipeline is rarely the problem. The problem is keeping up with it.
Each quote was taking the estimating team between six and ten hours of focused work. That meant pulling current prices from suppliers, taking off measurements from drawings, writing scope and exclusions, formatting the document so it looked the part, and double checking the maths before it went out the door. With two estimators and a working week that already includes site visits and client calls, the practical ceiling was around twelve quotes a month.
The owner had been watching the lost pipeline for some time. Tenders were coming in faster than the team could turn them around, and roughly one in three was being politely declined or quietly missed because there was no realistic way to respond inside the deadline. Some of those were the kind of jobs the business was perfectly placed to win. Watching that work go to competitors, knowing the team could have delivered it well, was the part that stung most.
He had been pitched plenty of estimating software over the years and had not bought any of it. The objection was always the same. He did not want a system that pretended to know more about pricing a job than the people who had been doing it for twenty years. He wanted his estimators in the driving seat. He just wanted them to spend their time on the parts that actually needed their judgement.
What we did
We sat with both estimators for two days and watched them quote a real job from start to finish. The pattern was clear. The hours were not going on the difficult bits. The hours were going on assembly, on hunting down supplier prices that had moved since the last quote, on retyping the same scope language for the tenth time, and on making the final document look consistent.
So that is what we built around. We set up a tool that pulls current rates from the merchants the business already uses, keeps a living library of the scope wording the team has refined over the years, and assembles a properly formatted draft quote from a short brief the estimator types in. The estimator points it at the drawings, gives it the job context, and gets back a draft with measurements pulled, materials priced at current rates, scope and exclusions written in the house style, and the document laid out the way clients expect to see it.
What we did not touch was the estimating itself. The numbers still belong to the estimator. They review every line, adjust the rates where their experience tells them the standard price is wrong for this client or this site, add the labour judgement that no tool can make for them, and sign the quote off. The tool does not guess at margin. It does not pretend to understand site conditions. It hands over a tidy, accurate starting point and gets out of the way.
The build took around three weeks from first conversation to the team using it on live tenders. We trained both estimators in person and stayed close for the first month while they pushed back on the bits that were not quite right.
The result
Quote turnaround dropped from six to ten hours to around 90 minutes. Monthly quote volume went from roughly twelve to twenty six. Win rate moved from 22% to 28% across the first quarter of use, partly because more quotes went out on time and partly because the team had time to tailor the ones that mattered most. The owner estimates the additional bidding capacity is worth about £1.4m in annual pipeline that the business simply could not have chased before.
When we asked the owner what he tells other contractors who ask about it, he said this. "I was ready not to like it. I have seen too many tools that try to be cleverer than the lads. This one just does the donkey work and lets them think. That is all I ever wanted."
Frequently asked questions
Will an AI estimating tool try to be cleverer than my estimators?
Not the way we set it up. The numbers still belong to your estimators. The tool pulls current rates from your merchants, retypes scope language from your library, and assembles a tidy draft document. The estimator reviews every line, adjusts where their experience tells them the standard price is wrong, and signs the quote off. The tool does the donkey work. The judgement stays with the people who have been doing it for twenty years.
How do you keep supplier prices current?
We connect to the merchants the business already uses, so the rates pulled into a draft quote reflect current pricing rather than whatever was true six months ago. Where a merchant does not have an integration, we fall back to a structured update process so the estimator can refresh the rates in seconds rather than chasing them on the phone. Either way the estimator sees the source of every number and can override it.
How long does it take to set up?
Three weeks from the first conversation to the team using it on live tenders is typical. About two days are spent watching the estimators quote a real job from start to finish so we understand the pattern. The rest is configuring the supplier connections, loading the scope-wording library, and tuning the document layout to match the contractor's house style. We then stay close for the first month to refine the bits that need it.
What does this kind of engagement cost?
Scoped per business. The variables are how many estimators, how many supplier integrations, and how much existing scope-wording library can be reused. Most contractors of this size pay for the engagement out of the extra pipeline they can chase in the first quarter. The free AI Opportunity Report puts honest numbers on the table before any commitment, with no obligation either way.