When an RFQ email arrives with PDF drawings, an AI agent reads the title block and notes for material, quantity, tolerances, finish and heat treatment, compares them with the shop's capabilities, drafts questions about gaps and queues clean requests with a feasibility summary.
Typical results
≈15 min
Saved per RFQ on reading drawings
≈7 hrs/wk
Estimator time back at around 30 RFQs a week
Same day
Spec questions sent back to the customer
Figures reflect typical results for this kind of automation.
The challenge
RFQs arrive with PDF drawings in every style, and the estimator reads each one for material, tolerances, finishes and heat treatment before pricing. Missing specs or an outsourced process often surface only after the quote is sent, so quotes get revised, margins slip and some jobs land on machines that cannot hold the tolerance.
Before and after
Minutes of drawing review per RFQ
−60%Quote revisions per month
−40%Hours to send spec questions to the customer
−92%The automation
When an RFQ email arrives, an AI agent reads each PDF drawing's title block and notes for material, quantity, tolerances, surface finish and heat treatment. It compares them with the shop's capability sheet of machines, part envelope and in-house processes. Missing or conflicting specs become a clarification email drafted for the estimator, and steps such as plating or heat treatment are flagged as outsourced. Clean requests land in the quoting queue with a one-page feasibility summary.
How it works
- 01
RFQ email arrives
An RFQ email with PDF drawings lands in the quotes inbox, and each drawing is picked up automatically.
- 02
Title block and notes read
An AI agent reads material, quantity, tolerances, surface finish and heat treatment from the title block and drawing notes.
- 03
Compared with shop capability
Each part is checked against the capability sheet: machines, part envelope, tightest holdable tolerance and in-house processes.
- 04
Gaps become a draft email
Missing or conflicting specs become a clarification email drafted for the estimator, and outsourced steps such as plating are flagged.
- 05
Clean RFQs queued to quote
Complete requests land in the quoting queue with a one-page feasibility summary for each part.
Starts from
- RFQ emails
- PDF part drawings
- Shop capability sheet
Delivers
- Clarification email draft
- Feasibility summary
- Quoting queue entry
The impact
Quote revisions drop by about 40%, because questions about specs and outside processes are settled before pricing. The estimator spends around 10 minutes per RFQ instead of 25 and sends clarifications the same day, and parts beyond the shop's envelope or tolerance are caught before a price goes out.
Questions about this automation
Does the AI set the price?
No. The agent reads the drawings and checks feasibility, and the estimator still prices every job. What changes is that the estimator starts with the specs already pulled out, the gaps already listed and the outside processes already flagged, so the time goes into pricing instead of hunting through drawing notes.
What if a drawing is a poor scan or has no title block?
The agent marks any field it could not read with confidence instead of guessing, and those RFQs go to the estimator with the unclear fields highlighted. If a drawing is unreadable, the clarification draft asks the customer for a cleaner PDF or a STEP model of the part.
What does the shop need to provide?
A capability sheet listing machines, travel and part envelope, achievable tolerances and finishes, plus which processes are done in-house and which go out. Access to the quotes inbox and the quoting queue completes the setup. This is a multi-step build, usually rolled out in stages, starting with drawing extraction.