Proposal Automation for an Industrial Supplier
A mechanical engineering supplier reduces proposal creation time from 4 hours to under 20 minutes — with an AI agent that reads enquiries, calculates costs, and generates a finished proposal in the ERP template.
Starting point
A mid-sized mechanical engineering supplier (approx. 80 employees) receives 15–30 enquiries per day for custom components. The sales team was spending 3–4 hours per proposal on manual cost calculation, material research, and formatting into a Word template.
The result: proposals arrived late, sales staff were frustrated, and orders were going to faster competitors.
Solution
Opteria built a three-stage AI agent:
- Enquiry parser – extracts structured data from unstructured emails and PDFs (dimensions, materials, quantities, delivery date)
- Calculation module – accesses material prices and production parameters from the ERP system and calculates costs + margin according to company rules
- Document generator – fills the company’s own proposal template (Word/PDF) and stores the document in the CRM
Human-in-the-loop: the sales rep reviews the generated proposal in under 5 minutes, adds optional comments, and clicks “Send”.
Technology
- Claude 3.5 Sonnet (document understanding, calculation)
- n8n (workflow orchestration)
- ERP integration via REST API
- HubSpot CRM (storage & dispatch)
Results
| Metric | Before | After |
|---|---|---|
| Creation time / proposal | 3–4 h | 18 min |
| Proposals / week | 25 | 75 |
| Callback due to errors | ~8 % | < 1 % |
| Sales team satisfaction (1–5) | 2.8 | 4.6 |
“We’re now sending proposals before the competition has even started.” — Head of Sales, anonymised
Investment
Project scope: Pro tier (€ 6,500 incl. ERP integration). ROI reached after approx. 6 weeks.
Ready to implement AI in production?
We analyse your process and show you in 30 minutes which workflow delivers the highest ROI.