Introduction

AI-enhanced fault-finding tool is an AI assistant for facility management (FM) technicians that uses a camera, Vision/OCR, and ChatGPT/Claude/Gemini intelligence to identify assets, detect issues, and guide fault resolution on-site.

Proof of Concept Goals and Coverage

Proof of Concept Goals and Coverage

Original Goal

How the POC Covers It

Technician uses a mobile camera on-siteThe Expo mobile app lets the user pick or capture a photo from their phone.
Recognize the asset/component in the photoMock vision service in the orchestrator simulates object detection + OCR (swap for real Vision API later).
Send image/context to ChatGPT backendThe orchestrator builds the Object Analysis prompt from our prompt library and sends it to ChatGPT.
ChatGPT performs the "review"ChatGPT identifies the object, adds function, safety, PM tasks, common faults, and spare parts info in structured JSON.
Return actionable response to technicianThe mobile app displays all the AI’s findings, including when confidence is low and further images are needed.
Fault finding guidanceThe same structure supports plugging in the Troubleshooting Guided Flow prompt for step-by-step fault diagnosis.
Backend integrationCMMS mock endpoint in the orchestrator shows how logs could be created automatically in your work order system.

Was this page helpful?