Crowd Sense

Every day, millions of citizens in Indian cities face broken roads, overflowing garbage, failed streetlights, and waterlogging — yet most complaints never reach the right authority, get lost in WhatsApp groups, or are filed manually with no location, no priority, and no follow-up. The civic system doesn't fail because officials don't care. It fails because the information reaching them is scattered, unstructured, and impossible to act on at scale.

In Bhopal alone, the municipal corporation manages 85 wards across 7 zones with hundreds of field officers — yet there's no unified, intelligent system that tells them what is broken, where exactly, how urgent it is, and who should fix it.

Citizens don't know which department owns a pothole. Officers don't know which complaints are critical. Duplicate reports pile up. SLAs get missed silently. Hotspots go unnoticed until they become crises.

We're solving the last-mile gap between a citizen's frustration and a government officer's action.

CrowdSense AI is a real-time civic complaint intelligence platform that transforms a citizen's phone photo into a fully triaged, department-routed, and publicly trackable civic complaint — in under one minute.

A citizen snaps a photo of a pothole, garbage pile, broken streetlight, or waterlogging. The app automatically captures their GPS location, detects the exact Bhopal Municipal Corporation ward and zone using real polygon geofencing (not guesswork), and reverse-geocodes the address. Anthropic Claude then classifies the complaint — assigning category, severity score, civic impact score, risk tags (public health, safety, environmental, traffic), and a recommended action — all from the photo and a short description. A deterministic routing engine (never AI) assigns the complaint to the correct BMC department and sets an SLA deadline.

But CrowdSense AI doesn't stop at reporting. It runs duplicate and cluster detection — if a similar complaint already exists nearby, citizens are shown a match percentage and invited to support the existing issue instead of creating noise. Repeat reports form hotspot clusters visible on a live map.

On the admin side, BMC officers get a civic command-center dashboard with a three-tier role hierarchy — Super Admin, Zone Senior Officers, and Field Officers. Each officer sees only what they're responsible for. SLA breaches trigger automatic escalation up the chain. An AI-generated briefing narrates the live state of the city — open complaints, hotspots, department performance, and emerging trends — over real, already-computed data.

The result: Citizens get accountability. Officers get clarity. The city gets a data-driven feedback loop between public complaints and government resolution — built on real Bhopal ward and zone geography, powered by Claude, and deployed for the Claude Impact Lab.

Claude code

Anthropic Claude then classifies the complaint — assigning category, severity score, civic impact score, risk tags (public health, safety, environmental, traffic), and a recommended action — all from the photo and a short description. The city gets a data-driven feedback loop between public complaints and government resolution — built on real Bhopal ward and zone geography, powered by Claude, and deployed for the Claude Impact Lab.