Can ChatGPT or Claude Do Construction Estimating? What Works and What Doesn't
Plain ChatGPT and Claude guess at quantities and rates. Set up as a 15-step process with your rules and rates, they produce a BOQ you can check.
Almost every subcontractor we talk to has tried it. Drop a drawing into ChatGPT, ask for a takeoff, get a confident table back. Check a few numbers and they’re wrong.
That doesn’t mean AI can’t estimate. Estimating was never one prompt. A good estimator works through steps, each with rules. Give the AI those steps and rules and the result changes.
Why does a single prompt get it wrong?
Paste a drawing and ask for a price, and the AI guesses everything that matters:
- Scale. Read it wrong and every length is wrong, silently.
- Scope. It doesn’t know which sections you price and which are by others.
- Allowances. Cable per point, access equipment, testing and commissioning aren’t drawn. Your rules hold them.
- Rates. It fills in a plausible price. Not your price.
- Proof. A number with no way to check it.
So you redo the takeoff to check it. That’s the work you were trying to save.
What actually works?
We set it up the way a careful estimator works: 15 steps in four stages.
- Drawings to quantities (steps 1–7). Register every file with its revision. Agree scope against your usual sections. Check the scale on every drawing. Count every fitting, recording where each count came from. Measure the runs. Add your allowances. Fill the BOQ in your own structure.
- You confirm the BOQ (step 8). Every row has a drawing screenshot beside it. Anything unsettled waits on one checkpoint tab.
- Pricing (steps 9–13). Draft supplier RFQs (you press send). Load your rates and labour; missing or expired rates are flagged, never filled. Work out direct costs, apply mark-ups and GST. You approve the price.
- Quotes and changes (steps 14–15). One quote letter per builder on your letterhead, each with its own revision. An addendum re-measures and re-prices only what changed.
Sample data. The estimator checks each line against the picture beside it: number, picture, number, picture.
In ChatGPT, the estimator runs one command on the tender folder:
Sample demonstration in ChatGPT.
It checks its own counts and says what it can’t settle:
Sample demonstration in ChatGPT. Counts are checked against the drawing, and gaps are listed.
You come in three times: confirm the BOQ, approve the price, approve each letter. Nothing is sent without you.
What’s the difference in practice?
| One prompt in ChatGPT | Set up as a process | |
|---|---|---|
| Where a quantity came from | Not shown | Drawing, page and revision on every line |
| Scale | Assumed | Checked on every drawing |
| Things it can’t read | Guessed | Flagged on one checkpoint tab |
| Rates | Generic | Your rate register, with source and date |
| Your estimator | Redoes the takeoff | Checks and signs off |
| Addenda | Start again | Change schedule with the price difference |
The third row matters most. A setup that tells you what it couldn’t settle is one you can trust with the rest.
Sample demonstration in ChatGPT. Unresolved quantities stop the job for a person, not a guess.
Where do the prices come from?
Pricing is where generic AI does the most damage. We load your rate register, each rate with a source and date, then your labour hours and mark-ups.
Sample data. Each rate carries its source and date. Missing rates are listed for you to fill in.
You get an internal pricing workbook with costs and margins, and a separate draft quote on your letterhead without them. Nobody types the quote out by hand again.
Sample data. The customer copy never shows your internal costs or mark-ups.
Should I use ChatGPT or Claude?
Both run this process. Our plugin works in either, so use the one your business already has. What matters more: it runs in your own business account, so drawings and rates stay in your workspace.
So, can ChatGPT or Claude do your estimating?
On their own, not for numbers you’d sign. Set up around your rules, rates and your estimator’s sign-off, yes. It takes the counting, typing and repricing off your plate. That’s what we set up: see how our estimating plugin works.
For what a good setup should do, read the best AI estimating setup for Australian subcontractors. For electrical work, see electrical estimating software.
Frequently asked questions
Can ChatGPT read construction drawings?
It reads text, legends and schedules and counts symbols. On its own it can't reliably work out scale, decide your scope or know what it missed. That needs rules and checks.
Is Claude or ChatGPT better for estimating?
Both run the same process once it's set up. Our plugin runs in either, so pick the one your business already pays for.
Will AI replace my estimator?
No. It takes the counting, typing and repricing. Your estimator confirms quantities, makes the scope calls and approves the price.
Where do my drawings and rates go?
They stay in your own business Claude or ChatGPT account and your own storage. On a business plan, your data isn't used for training by default.
Want this on your own jobs?
The estimating plugin is live today: we set it up in your own Claude or ChatGPT account around your rates and your template, and your estimator checks every line before it goes out. The App, with the rest of the office, is in build.