AI quote automation for manufacturers: how it works and where it breaks

Quoting is one of the clearest places for a manufacturer to try AI, because the work is repetitive, the inputs are documents, and the cost of delay is easy to see. A request for quote (RFQ) arrives by email with drawings attached. Someone has to read it, find comparable past jobs, check material and routing, and produce a price and lead time. When the estimator is busy, the RFQ waits.

How AI-assisted quoting works

• Intake: the RFQ email and attachments are captured automatically and logged, so nothing sits unread.
• Extraction: the system pulls out the structured facts: part number or description, quantity, material, finish, tolerances called out, due date and customer.
• Lookup: it searches your history for similar parts and pulls past quotes, routings and material costs from the ERP or a quote spreadsheet.
• Draft: it proposes a price and lead time and shows the past jobs it used, so the estimator can see why.
• Review: the estimator edits and approves. The system does not send a price on its own.
• Learn: the outcome (won, lost, actual hours) is recorded so the next draft is better grounded.

Where it breaks

• Drawings with missing or unclear tolerances, where a person has to ask the customer.
• New part families with no history to compare against.
• Volatile material prices, which need a live or recently updated source.
• Customer-specific terms, rush rules and minimums that live in someone's head.
• Stale ERP data: routings and standard hours that no longer match the shop floor.

Most of these are data and process problems, not model problems, which is why an assessment of your quote data comes before any build.

What to measure before and after

• Time from RFQ received to quote sent.
• Share of RFQs that receive a quote at all.
• Estimator hours per quote.
• Win rate, and the gap between quoted and actual margin.

Measure these for two to four weeks before you change anything. Without a baseline you cannot tell whether the system helped.

Context on adoption

The U.S. Census Bureau's Business Trends and Outlook Survey found that, as of May 3, 2026, 19.8% of U.S. businesses overall reported using AI, compared with 37% of firms with at least 250 employees and 32% of firms with 100 to 249 employees. Adoption rises with firm size, which means a mid-size manufacturer that starts with one narrow workflow is not behind a crowd; it is early. McKinsey's 2025 survey adds a caution: among organizations using AI agents, no more than 10% say they are scaling them in any single business function, so most deployments are still young.

Sources

• U.S. Census Bureau, "Large Firms With at Least 20 Employees Biggest AI Users" (May 2026)
• McKinsey, "The State of AI: Global Survey 2025"