background
A real estate company had to analyze several thousand energy bills (electricity, gas, water) per year to monitor consumption by building and by meter. PDF documents, which were unstructured and heterogeneous, made reliable and rapid analysis impossible.
Challenge
- PDF invoices in variable formats depending on the supplier
- Manual data extraction (dates, amounts, counters...)
- Complex reconciliation with the accounting base
- Multisite monitoring that is difficult to consolidate
- No tool to visualize or alert on discrepancies
AI solution
Strat37 deployed a dedicated AI agent capable of:
- Automatically convert PDF invoices into usable text (OCR)
- Identify the type of bill (energy, services, etc.)
- Extract key fields: amount, dates, consumption, counter, supplier
- Enrich the data with the existing accounting database (site, analytical account...)
- Save everything in a unified database (MongoDB)
- Show data in StratBoard™ with filters by building, meter, and period
- Integrate a collaborative process with continuous feedback on the quality of scanned documents
Results
✅ +1,000 invoices analyzed automatically
✅ 100% of data reconciled with accounting
✅ Classification into +70 service categories
✅ Consolidated monitoring of consumption by meter and by site
✅ Data ready for strategic analysis and multi-site generalization