▫️ Built on:
▫️ LangChain + custom OCR for extracting and understanding technical documentation
▫️ OpenAI Whisper + TTS for 12-language voice input/output
▫️ Real-time visual part recognition using smartphone cameras
▫️ CRM + ticketing integration for seamless handoff when needed
▫️ The Results:
▫️ 📉 52% fewer support calls
▫️ 🌍 Language coverage grew from 3 → 12
▫️ ⏱️ Avg troubleshooting time dropped 37%
▫️ ✅ First-time fix rate jumped from 62% → 84%
▫️ 💰 $1.2M annual support savings (projected)
▫️ Inside the System:
▫️ Voice-led, Multilingual Support
▫️ Technicians speak naturally in their language
▫️ Assistant guides them step-by-step via voice, no scrolling
▫️ Text-to-speech answers tailored to machine model, error code, and user’s location
▫️ Smarter, Safer Troubleshooting
▫️ Recognizes part numbers and CAD diagrams from camera input
▫️ Warns of safety issues in real-time
▫️ Tracks progress through complex repairs
▫️ Seamless Escalation—Only When Needed
▫️ If unresolved, the system creates a detailed support ticket
▫️ Auto-attaches transcript + visual logs
▫️ Routes to right tier of support based on complexity