A request for quote lands at 4pm on a Thursday. It is a PDF with six pages of specifications, two drawings and a reference to last year's schedule of rates. The estimator opens it, reads it, pulls up three similar jobs from the last two years to check what they actually cost, adjusts for the new scope, types the line items into the quoting spreadsheet, writes the covering email and sends it by Friday. If the business quotes ten jobs a week, that estimator is spending most of their time on reading, searching, copying and formatting. The thinking, the bit where experience tells them that job is going to be tighter than it looks, takes a fraction of the time.
That reading-searching-copying part is where AI genuinely helps. Not by replacing the estimator, but by doing a first pass on the repetitive work so the estimator spends their time on the judgement calls.
What AI can do well in quoting
AI is good at reading and extracting. Give it a PDF of a specification, a drawing legend, a request for quote or a supplier price list, and it can pull out the structured information: line items, quantities, dimensions, material types, dates. Not perfectly every time, but well enough to save the estimator from retyping it. We covered the broader picture of AI reading documents in AI for dockets, orders and supplier invoices.
Here is where it fits in a quoting workflow:
Reading the incoming RFQ or spec. AI can parse a PDF or email, pull out the key requirements (items, quantities, material grades, delivery dates, site details) and present them in a structured format the estimator can check against. Instead of reading six pages and taking notes, the estimator reviews a summary and confirms or corrects it.
Searching past jobs. This is one of the most useful applications, and one most businesses do not think of. If your quoting system or job records are searchable, AI can find the three most similar past jobs based on scope, materials, size or customer, and pull up what they were quoted, what they actually cost, and where the margin sat. The estimator still decides which past job is the best comparison, but the search takes seconds instead of half an hour of scrolling through folders.
Pulling current prices. If supplier price lists are in PDFs or spreadsheets (and they usually are), AI can read them and match items to the line items in the quote. This is particularly useful when you deal with dozens of suppliers whose price lists arrive in different formats. The estimator checks the prices rather than looking each one up.
Drafting the quote document. Once the estimator has confirmed the line items, quantities and prices, AI can assemble the quote in your standard format: the covering letter, the schedule of rates, the inclusions and exclusions, the terms. The estimator reviews and adjusts, but the assembly is done.
Drafting the covering email. A small thing, but it adds up across fifty quotes a month. AI writes a professional response referencing the RFQ number, the scope summary and the key terms. The estimator reads it, adjusts the tone if needed, and sends it.
Where AI struggles, and where a person stays in charge
AI does not understand your business the way an experienced estimator does. It does not know that the site access on that particular project is going to add two days, or that the client always changes the scope after the quote is accepted, or that the steel price is about to jump because of a tariff announcement.
Judgement on risk and margin. AI can calculate a cost based on quantities and rates. It cannot assess whether to add a contingency because the drawings look incomplete, or whether to sharpen the price because this client is worth winning. That is the estimator's job, and it is the part that actually decides whether the business makes money.
Understanding ambiguous specs. A specification that says "or equivalent" or "to engineer's approval" means something specific in context. AI reads the words but does not always understand the commercial implication, which might be that the client is open to an alternative material, or that they are reserving the right to reject your work later.
Checking the final number. Before a quote goes out, someone who understands the job should look at the total and ask whether it makes sense. AI can help with the arithmetic, but the sense-check (does this margin look right for this kind of work, this client, this market?) is human.
Handling drawings and plans. AI's ability to read technical drawings is improving but not reliable enough to trust without checking. It can sometimes extract dimensions and quantities from simple drawings. For complex engineering or architectural plans, the estimator still needs to do the takeoff, or use a dedicated takeoff tool. AI can help label and organise what comes out.
The rule we apply with every AI project is that a person checks and approves the output before it goes anywhere. For quoting, that means the estimator reviews every AI-generated summary, price lookup and draft quote before it reaches the customer. AI does the first pass; the estimator does the thinking and the sign-off. We wrote about this approach more broadly in what not to use AI for in your business.
What the workflow looks like in practice
Here is a realistic version of the Thursday afternoon RFQ, with AI handling the first pass:
- RFQ arrives by email. AI reads the attachment, extracts the scope items, quantities, material specs and dates, and presents a structured summary.
- Estimator reviews the summary. Confirms or corrects the extracted information. Adds context the AI cannot see: site difficulty, client history, risk notes.
- AI searches past jobs. Finds the three closest matches in the job system or quoting database, showing quoted price, actual cost and margin for each.
- Estimator selects the best comparisons and adjusts for the new scope: different quantities, different materials, different site conditions.
- AI pulls current supplier prices for the main materials and presents them against the rates from past jobs.
- Estimator builds the price, adding their margin, contingencies and any scope adjustments.
- AI assembles the quote document in the company's standard format, with the covering email.
- Estimator reviews, adjusts and sends.
The estimator still does every step that requires experience and judgement. AI handles the reading, searching, matching and formatting. In a business that quotes ten or more jobs a week, that saves real hours, and it means the estimator can spend more time on the quotes that matter most.
What you need before AI can help with quoting
AI is not magic, and it cannot work with data it does not have. Before you can use it for quoting, you need:
- Past job records in a searchable form. If your old quotes and job costs live in folders of PDFs, Word documents and spreadsheets with no consistent naming, AI cannot search them usefully. The first step might be getting your quoting history into a structured format, even a well-organised spreadsheet.
- Supplier price lists that AI can read. Most price lists arrive as PDFs. AI reads PDFs well, but if the format changes with every supplier, it takes some setup to teach it each layout. Standard formats help.
- A quoting template. AI needs to know what your quote looks like: the sections, the terms, the format. A consistent template means it can assemble the output without reinventing it each time.
- Clean data in your systems. If your job system has item codes and descriptions that mean different things to different people, AI's past-job search will return noise. The same data quality issues that affect every integration apply here.
If your systems are not there yet, the answer might be to fix the data and processes first, then add AI. We have seen businesses get more value from organising their quoting records than from any AI tool. Our AI readiness checklist walks through what to check before you invest.
Frequently asked questions
Can AI write a quote for me?
AI can draft a quote document based on the line items, quantities and prices you confirm. It cannot decide the price, assess the risk or make the commercial judgement. Think of it as a very fast assistant who reads, searches and formats, but always needs the estimator to check and approve.
Which industries benefit most from AI in quoting?
Any business that quotes frequently from written specifications or requests: fabrication, wholesale supply, trades, manufacturing, engineering services. The more standardised the quoting format and the more repetitive the reading work, the more time AI saves.
Does AI replace estimators?
No. It handles the repetitive parts of the job (reading specs, searching past jobs, pulling prices, assembling documents) so the estimator can focus on the parts that need experience and judgement. A good estimator with AI support can handle more quotes, or spend more time on the ones that need care.
How accurate is AI at reading specifications and RFQs?
It depends on the document. Clean, well-structured PDFs with consistent formatting are read well. Scanned documents, handwritten notes, complex tables and ambiguous language need more checking. The estimator always reviews the extracted information before using it.
What does setting this up involve?
It depends on what you have now. If your quoting records are well-organised and your price lists are in a consistent format, a pilot can start small. If the first step is organising your data, that comes before AI. We scope each project after understanding your current systems and quoting process. Take our AI fit check for a quick sense of whether your setup is ready.
Next steps
If your estimators spend more time reading and searching than thinking, and you want to know what AI could realistically do for your quoting process, book a free initial chat. We will look at your current workflow and tell you honestly whether AI makes sense now or whether something else should come first. You might also find these useful: