In a job shop, the quote is built from information that lives in several places, so the first job for AI and automation is to pull that information together for the person quoting, one RFQ at a time. That person still decides the price, and a person approves every quote.
The case for that order comes from one Modern Machine Shop article on quoting, a column on AI in manufacturing, and two federal and association sources on the workforce.
The Quote Feeds the Rest of the Shop
A customer sends a request for quote, or RFQ, and the shop answers with a quote for the part. A Modern Machine Shop article by Julia Hider, published February 23, 2026, quotes Francois Gau, founder and CEO of the industrial marketing firm GrowthHive: "Quoting is the heart of the operation. This is what feeds the rest of the business."
O*NET OnLine describes the cost estimator's work as preparing cost estimates for product manufacturing, construction projects, or services, to help management bid on or price a product or service. Two of its listed tasks are analyzing blueprints and other documentation to prepare time, cost, materials, and labor estimates, and collecting historical cost data to estimate costs for current or future products. That occupation also covers construction and services, so read it as a description of the work, not a count of job shop estimators.
The Inputs Sit in Different Systems
Gau says a quoting team has to "make very quick decisions with a host of information that's scattered around your systems." The article lists machine productivity, cycle times, changeover time, tool life, inspection reports, and outside processing time, and says there is more.
He also says some ERP programs can generate quotes, but might not have the details accurate quoting needs, or just are not dynamic enough. His example: a shop with several models of the same machine on its floor often has no ERP record of exactly which machine made a part or how many hours it spent on it. The article adds that shops can build spreadsheets or invest in third-party quoting software.
This is where job shop quoting automation earns its place, in our reading. For each RFQ, a system can pull what the shop already records into one view, show which inputs are missing, and hand the person quoting a complete picture. Setting the price is a different job, and it stays with a person.
Gau is blunt about the choice of tool: "There's so much variation it's almost impossible to provide a general conclusion as to which automation solution is going to be most efficient."
Speed Helps, and Accuracy Decides the Job
According to Gau, the new industrial buyers prioritize speed, and he says "Automation is a key tenet because we have hours, not weeks" to turn a quote around. In the same article he cautions that speed has to be balanced with accuracy. Missing a callout for a process might mean the difference between a job that makes money and one that loses money.
So a faster quote that leaves out an input is worse than a slower one that has it. Gau's suggested starting point is simple:
- Appoint a quoting champion, whose first task is to establish the current process. - Trace the flow of information and material from receiving an RFQ to finished product. - Let the champion find the bottlenecks and take them to shop management.
Do that before choosing a tool. The path you trace is what any system will have to follow.
What Stays With a Person, and What Stays Out of Chatbots
Gau sees applications for AI in manufacturing, particularly in the front office. He still encourages caution about AI chatbots for quoting, saying, "The limitations are more business risks and IP mitigation than the capabilities." His advice is that shops should not put customer names or any information about the parts they are quoting into chatbots. Before connecting any tool to your quotes, find out where that information goes.
A Modern Machine Shop column by Brent Donaldson, dated August 13, 2025, describes how AI differs from CNC. CNC is deterministic: as long as the inputs and setups stay the same, machinists can trust it to produce the same motions and results every time. AI operates differently. It needs clean data to learn from patterns and make suggestions, and it asks us to trust a system we cannot really see. The column also notes that AI has not been widely adopted by the magazine's core audience.
That difference is why we recommend this setup: AI assembles and flags, and a person reviews and approves. Clean inputs come first.
Who Holds the Knowledge
Some context on the workforce, with the scope stated. O*NET OnLine reports 288,300 machinists employed in 2025, with projected growth of 1% to 2% for 2025 to 2035, slower than average. That counts the whole occupation across all industries, not job shops. The Manufacturing Institute, with Deloitte, projects that as many as 3.8 million additional employees could be needed in manufacturing from 2024 to 2033, and that 1.9 million of those jobs could go unfilled if workforce challenges are not addressed. Those are manufacturing-wide projections, not job shop data.
Neither source says anything about quoting. The practical point is ours: if knowing where the inputs live depends on one person's memory, write the path down while that person is still there to explain it.
Where to Start
Pick one part you have quoted recently. List every piece of information in the quote and where each one comes from: a machine record, a person, a spreadsheet, an inspection report, an outside processor. Mark the ones that are hard to find or missing. Fix that path first, then decide what a system should gather.
Our guide to business software integration covers getting the systems you already own to share information, and our article on office efficienc-automation) covers the front-office tasks around it.
To map one quote from RFQ to finished part, get your free AI assessment.
Frequently Asked Questions
Can AI set the price on a job shop quote?
We do not recommend it. Gau's caution is that a missed process callout can turn a money-making job into a losing one, so a person should review and approve every quote.
Is it safe to put customer names and part information into an AI chatbot?
Gau advises against it. He says shops should not put customer names or any information related to the parts they are quoting into chatbots. This is not legal or compliance advice.
Does the ERP already hold what a quote needs?
Not necessarily. Gau says some ERP programs might not have the details needed for accurate quoting, such as which machine made a part and for how many hours. Check yours against your own list of inputs.
Where should a job shop start with automation?
With one recent quote. List where each input comes from, then fix the path before buying anything.
