The future of estimating for UK trades

Estimating has changed less in fifty years than almost any other part of running a building firm. The tools went from paper to spreadsheets, but the craft, one experienced person turning drawings into a price, has stayed stubbornly manual. That is now changing quickly, and the firms that understand the direction of travel will spend the next decade winning work from the ones that do not.

How we got here: three eras of the same craft

The first era was paper: scale rules, dim sheets, and rates carried in a black book and a good memory. The second era was the spreadsheet, which made the arithmetic faster but left everything else untouched: the estimator still measured by hand, still retyped rates, still worked evenings, and the firm's pricing knowledge still lived in one person. Most UK trade firms are in this second era right now, and its limits are familiar: quotes take days, consistency depends on who does them, and when the estimator is on holiday the pipeline stops.

The third era, arriving now, is structured pricing: the job is described once, in words or drawings, and software builds the measured, itemised bill from it, using the firm's own rates. The human moves from typing to judging. That shift sounds small. It is not, because it changes what an estimator's hour is spent on, and how many quotes a firm can produce with the people it already has.

What is genuinely new

Three capabilities matured recently and together they explain the moment. First, language models can now read a plain-English job description, or a set of architect's drawings, and extract the scope with useful reliability. The honest limits of that are worth understanding, and we set them out in can AI really read construction drawings. Second, deterministic pricing engines can turn that extracted scope into a measured bill of quantities, the same input producing the same output every time, which is what makes the result auditable rather than a chatbot guess. Third, the tenant rate book: the software prices with your day rates, your material costs and your regional reality, not a national average pretending to be everyone.

The important design point, and the reason serious tools keep a human in charge, is that the machine produces a draft, not a decision. Every line stays editable. The estimator's judgement about access, ground, client and risk is still what turns the draft into a price. That division of labour is the whole architecture of trustworthy AI estimating, argued fully in human in the loop estimating.

What it changes commercially

The obvious gain is speed, and speed is not a vanity metric: the first credible quote through the door wins a disproportionate share of domestic work. The subtler gains compound over time:

• Volume without headcount. A firm that could quote five jobs a week can quote twenty, which changes what marketing is worth doing.
• Consistency. The same job priced the same way regardless of who was tired that evening. Consistency is what lets a firm learn from its own numbers.
• A pricing asset. Every quote produced through a structured system enriches the firm's own rate book and win-rate data. The black book stops retiring when its owner does.
• Presentation as standard. Itemised, professional documents by default, which read as competence before anyone has met your work.

What to watch, honestly

A clear-eyed view includes the risks. Tools that promise a price with no visible quantities are asking for blind trust and do not deserve it: if you cannot see the lines, you cannot check the thinking. Firms that adopt speed without review discipline will send bad quotes faster than ever. And any AI that reads drawings will sometimes misread one, which is precisely why the workflow must surface what it measured for a human to challenge. The right questions to ask any vendor, including us, are the ones in trusting software with your margin.

One more thing worth watching is who owns the pricing data. A decade of your rates, your quotes and your win history is a genuinely valuable asset, arguably the most valuable thing a small firm's office produces. Before committing to any platform, ask how you get that data out, in what format, and at what cost. The answer tells you whether you are buying a tool or renting a dependency.

What a sensible firm does this year

You do not need a strategy document. You need a controlled experiment. Take one real job you have already priced, run it through a structured estimating tool at your own rates, and compare the bill line by line with what you sent. Where the tool missed, note it. Where you missed, note that more carefully, because omissions are where margins actually die. Then decide with evidence. This is the exact experiment QS Quoter is built to make cheap: it prices a full bill of quantities from a description or drawings at your rates, produces a client copy and a private cost copy, keeps every line editable, and the first quote is free.

For firms that want to go further, treating estimating as the first step of a wider systemisation of the office, that is a conversation about your whole workflow rather than one tool, and it is the work our commercial team does every week.