Service Centre
Intake Bot
Equipment intake via Telegram — no forms, no manual entry. A technician photographs the manufacturer plate, the bot reads it with OCR, extracts the key data, confirms with the operator, and pushes a structured record to ERP.
70%
Faster intake
0
Manual data entry
Any
Device type supported
// context
Problem & Solution
❌ Before
- Technicians manually typed device details into forms — slow and error-prone
- Serial numbers misread or mistyped led to wrong records in ERP
- Intake queues built up during busy drop-off hours
- No standardised process — every technician did it differently
✅ After
- Technician photographs the plate — bot does the rest
- OCR extracts manufacturer, model, and serial number automatically
- Operator confirms or corrects before data hits ERP — human-in-the-loop
- Structured JSON pushed to ERP instantly — no re-entry, no errors
// demo
Bot in Action
// features
Key Features
Photo-based intake
Technician sends a photo of the manufacturer plate. No app installation, no form to fill — just a photo in Telegram.
OCR extraction
Extracts manufacturer, model name, model number, and serial number from the plate image. Works on worn, angled, or partially obscured plates.
Human confirmation loop
Before writing to ERP, the operator confirms the extracted data — or edits it inline. Nothing gets saved without explicit approval.
ERP integration
Confirmed data is pushed as structured JSON to the ERP system. Compatible with any ERP that exposes a REST API.
Any device type
Works across PCs, laptops, printers, home appliances, and any other electronics with a manufacturer plate.
Full intake log
Every intake is timestamped and stored. The service centre has a full history of received equipment with operator confirmation records.
// outcome
Result
Equipment intake 70% faster — zero data entry errors
Technicians stopped typing device details manually. A photo replaces a form — the bot reads, the operator confirms, the ERP records. Intake queues during busy hours dropped significantly, and data quality in the system improved immediately after rollout.
// tech stack
Built With
Python
aiogram 3.x
OpenAI Vision
FastAPI