EngineeringVerified Real deployment at a named or precisely described company, with a measured result from a primary source.

AI agents turn RFQ emails and long customer specifications into a matched, priced quote in minutes (IMS Gear, Maschinenfabrik Reinhausen, a power-grid components maker)

Added 2026-11-09 · Last updated 2026-11-09

Manufacturers in this case: IMS Gear · Maschinenfabrik Reinhausen · all manufacturers using AI

Role

Sales engineering, application engineering and inside sales (quoting for configured or engineered products) · Typical users: sales engineer, application engineer, inside sales or quoting team, sales manager at an engineer-to-order or configure-to-order manufacturer

The problem

For manufacturers that sell configured or engineered products, a request for quotation arrives as an email with attachments, often a long specification, and someone has to read it, work out what is really being asked, match it to products, price it and assemble the quote. At IMS Gear, a quote ran through three departments by hand and took weeks. At Maschinenfabrik Reinhausen (MR), a global maker of transformer control and regulation technology, key information had to be extracted by hand from specifications with varying structures: slow, error-prone, and the interpretation depended on each person's expertise and region; a basic analysis of a 70-page document took about five hours by hand. A private-equity-backed power-grid components manufacturer received about 200 RFQ emails a month, all processed manually by inside sales.

How AI helps

A pipeline of AI agents reads the incoming RFQ, extracts the requirements, matches them to the company's products and drafts the quote, and people review what the agent is unsure about. At IMS Gear, a team of specialist AI agents built by the company's own engineers on Synera's platform turns a customer request into a finished quote. At MR, appliedAI built a chatbot for questions on specification documents and an automated pipeline that extracts the relevant specification details into customer-specific formats and matches the requirements against selected MR products; it is used by MR's domain experts, and the next phase links it to the product catalogue and quoting tools for LLM-assisted quote generation. At the power-grid manufacturer, Sia Partners built a generative AI pipeline on a cloud data platform that ingests RFQ emails, parses documents, matches products, calculates pricing and assembles the quote, with 55% of RFQs needing no manual intervention.

Step by step

  1. Collect 20 to 30 past RFQs with the quote you actually sent, including the awkward ones (long specifications, several products, missing information).
  2. Write down how a good quote is made: which fields you need, how you match a request to a product or option, the price rules, and who must approve what.
  3. Have the AI extract the requirements from each RFQ into a fixed list of fields, with the page or sentence each one came from, and compare with your quote.
  4. Add product matching against your catalogue, with a reason for each match and a flag for anything it is unsure about or any requirement you cannot meet.
  5. Keep pricing in your price rules or CPQ, and let the AI only fill in the inputs; have a person approve every quote at first.
  6. Track the share of RFQs that needed no correction, the time per quote and the errors found by the reviewer, then widen to more product lines.

Tools

  • Synera (platform for engineering AI agents): used at IMS Gear
  • appliedAI specification extraction and product-matching pipeline with chatbot: used at Maschinenfabrik Reinhausen
  • Sia Partners generative AI RFQ pipeline on a cloud data platform: used at the power-grid components manufacturer
  • Other tools to evaluate: the CPQ (configure, price, quote) module of your ERP or CRM, the agent features of your quoting software
  • To test the idea without new software: a few past RFQs, your product list and price rules in a spreadsheet, plus any LLM (ChatGPT, Claude, Copilot, Gemini) with the prompt below

In the AI tools index: appliedAI · Sia Partners RFQ pipeline · Synera · all AI tools used by manufacturers

Result

IMS Gear, Synera: time to quote down to 10 minutes from multiple weeks, more than 100 quotes processed in production, and one engineer instead of two per RFQ at the same volume (Synera customer story, a vendor publication that does not say how the earlier time was measured). Maschinenfabrik Reinhausen, appliedAI (six-month project in two phases, timeline given as Nov 24 to Nov 25): extraction of a 70-page specification in about 12 seconds against about 5 hours by hand, product matching of that document against 3 products in about 13 seconds against an estimated 3 hours, and a Q&A answer in about 3 seconds against 134 seconds; extraction accuracy was 90% for the basic analysis and 77% for the detailed analysis, and the solution is in production use by domain experts (appliedAI success story, a vendor publication; the manual times are the vendor's measurements or estimates). Power-grid components manufacturer (not named, about 200 RFQ emails a month), Sia Partners: standard quote time cut from 30 to 10 minutes (-67%), 55% of RFQs with no manual intervention, 96% email classification accuracy, 90% quote precision, 10% mean absolute percentage error on pricing, and an estimated $6.8 million in incremental annual revenue and $2.2 million EBITDA impact (Sia Partners, 4 September 2026; the revenue and EBITDA figures are estimates by the consultant).

As reported by the source; most figures are vendor or self-reported and not independently verified.

Human review: cautions

  • All figures are vendor or consultant publications. IMS Gear gives no baseline method; MR's manual times are estimates and its detailed-analysis accuracy is 77%; Sia's revenue and EBITDA effects are estimates for an unnamed client.
  • A wrong quote costs more than a slow one: keep price rules deterministic, keep a reviewer on every quote at first, and measure errors, not only speed.
  • These are engineered or configured products with structured catalogues; the approach needs clean product data and price rules to work.
  • Customer specifications and drawings can be confidential: check where the AI processes them before sending real customer documents.

Try it yourself

You are a sales engineer at a manufacturer of configured products, preparing quotes from RFQ emails.

I will paste: (1) one RFQ email with its attachment text [or a long customer specification]; (2) our product list with options and limits [codes, main specifications]; (3) our price rules or price list; (4) the quote we sent for a similar past request, if any.

Do this:
1. Extract the requirements into a table [requirement, value, unit, where it appears in the document] and list anything ambiguous or missing as questions for the customer.
2. Match each requirement to one of our products or options, say why, and mark any requirement we cannot meet or are unsure about.
3. List the inputs a price calculation needs, but do not compute the price: I will apply our price rules.
4. Draft the quote email in plain language with the open questions, ready for my review.
5. Tell me what a colleague should double-check before this goes out.

Never guess a technical value or a price. Use only the data I paste in; if something is missing, write 'not found' or list it as a question instead of guessing. Do not send, order, post or change anything in any system: produce drafts for my review and approval.

Want more like this? Join the newsletter

Related cases

All 13 Engineering use cases →

Sources

Weekly: 1–2 practical AI use cases for your plant

Newsletter in

Free, one email a week. By subscribing you agree to receive the weekly newsletter (and its sponsor message) from AI Agents for Manufacturing. Every email includes an unsubscribe link, and you can unsubscribe anytime. See our Privacy Policy.