pot.AI is an AI-powered public infrastructure monitoring system that turns everyday road cameras into continuous infrastructure inspectors.

It uses real-time computer vision to detect potholes, damaged roads, and broken streetlights from live camera feeds or uploaded road video. Each detection is captured as evidence and linked to real geographic location, timestamp, confidence, and detection history.

Gemma 4 adds a multimodal verification layer that analyzes detected evidence, helps validate the issue and severity, and provides a human-readable explanation for the decision. This creates a human-in-the-loop system rather than blindly trusting one model.

The platform tracks recurring infrastructure problems, prevents duplicate reports, prioritizes persistent issues, and supports a closed-loop workflow from detection and verification to escalation and eventual resolution.