The situation
Incoming quotations and purchase orders arrive as unstructured emails: PDFs, line items, delivery notes, and customer references all mixed together. Each order required someone to open the email, extract the details, find the right customer record, create the ERP entry, and send a confirmation.
At higher volumes, that process created slow turnaround times, avoidable address and contact errors, and unclear ownership across the inbox.
What we built
We rebuilt the intake flow as a single AI-powered pipeline. Incoming emails and attachments are automatically classified, order details are extracted, customer records are matched, and the ERP order is created without manual input.
Each order is assigned a clear status: processed, requires review, or error. Flagged orders go to a human with the relevant context already surfaced. Everything else flows through automatically.
The result
RFQ preparation went from 12 minutes to 2 minutes per request (83% faster). 80% of orders land straight in the ERP, ready to ship without human handoff. Every order carries a complete audit trail with the original email and attachments linked to it. The team now spends time on customer follow-up, exception handling, and order quality control instead of repetitive data entry.
As Allard, owner, put it: the automation saves around 10 minutes per RFQ and orders are processed significantly faster.
