An AI-powered CRM for stock, delivery notes, quotes and invoices — for a building-materials supplier
Based on projects we've delivered. Figures are approximate; client names are withheld for privacy.
The problem
Stock was in a notebook, quotes in Word, invoices in Excel. Nobody knew exactly what was left in the warehouse, or who was late paying.
Before and after
Before
Information scattered across four places.
After
One system that knows everything in real time.
What we built
- 1
Every stock movement, purchase order and delivery note is recorded in one system.
- 2
An AI agent prepares quotes and invoices from a single message and answers questions like 'How much cement is left?'
- 3
Automatic reminders for overdue invoices, and a dashboard that shows the situation in real time.
Results
- to prepare a quote, instead of 30
- 2 minto prepare a quote, instead of 30
- of delivery notes linked to their invoices
- 100%of delivery notes linked to their invoices
- returned to the office every month
- ~60 hoursreturned to the office every month
First month after launch compared with the month before.
Timeline
3 weeks
From first call to production
Under the hood
- n8n
- PostgreSQL
- Docker
- OpenAI / Claude
- Dashboard

Next project
RetailA WhatsApp agent that sells and answers for a fashion store
How we built it