Bengaluru Engineer Uses AI to Hunt Potholes and Expose the Contractors Behind Bad Roads: Here's How It Works

Bengaluru's pothole problem has become a familiar frustration for commuters, but a local engineer is attempting to tackle the issue using artificial intelligence and government data.

Engineer Gaurav Sen has developed a mobile application called Potholes Detector, which turns an ordinary smartphone into an AI-powered road inspection tool. The app can detect potholes while users are driving and, more importantly, connect those road defects to government road contracts and the contractors responsible for maintaining them.

The goal is to move beyond simply reporting potholes and create a system that can establish who is responsible for fixing them.

How Does the AI Pothole Detector App Work?

The application uses a smartphone camera to continuously monitor the road while a user is driving. The phone can be mounted on the vehicle's dashboard, allowing the camera to capture the road ahead.

The app uses an OpenAI vision model to analyse the captured footage and identify potholes. Detected potholes are categorised based on their size as small, medium or large.

The process is designed to be largely automated, meaning commuters do not need to manually photograph and record every pothole they encounter.

However, detection is only one part of the system.

AI App Connects Potholes to Government Road Contracts

The more interesting aspect of Potholes Detector is its database of public road contracts.

According to Sen, he spent months collecting and mapping municipal information covering more than 2,900 active government road contracts.

When the application detects a pothole, it uses the location coordinates to identify the corresponding road contract. It can then determine details such as the contractor and engineer responsible for that particular stretch of road.

The app can subsequently generate a complaint email containing the pothole's location, photographic evidence and relevant tender information.

This approach could make civic complaints more actionable by connecting physical road damage with the government contract responsible for maintaining that road.

Pothole Reports Could Help Enforce Road Warranties

Another important element is the defect liability period attached to many government road contracts.

Sen says many contracts include warranties of up to two years. If a pothole develops during this period, the responsible contractor may be required to repair the road without additional cost to the government.

By highlighting the relevant contract and its warranty period, the app could help citizens demand repairs while reducing the possibility of public funds being spent repeatedly on the same road.

How to Use the Potholes Detector App

The basic process is designed to allow commuters to detect and report road damage while travelling.

• Step 1: Mount Your Smartphone

Secure your smartphone on the vehicle's dashboard so that the camera has a clear view of the road ahead.

• Step 2: Start Driving Mode

Open the Potholes Detector app and activate its driving mode before beginning your journey.

• Step 3: Let the AI Scan the Road

As you drive, the smartphone camera continuously records the road surface. The AI analyses the footage and identifies potential potholes.

• Step 4: Allow Location Matching

When a pothole is detected, the application uses the vehicle's location coordinates to identify the relevant road section.

• Step 5: Check the Contract Details

The app cross-references the location with its government contract database to identify the contractor and responsible engineer, where matching data is available.

• Step 6: Generate the Complaint

The application can prepare a complaint containing the pothole's location, photographic evidence and relevant road contract information.

• Step 7: Send the Report to Authorities

Users can then send the structured complaint to the appropriate civic authorities, creating a documented report that can be followed up.

What Inspired the Bengaluru Pothole Detector?

The project was reportedly inspired by real-world accidents caused by Bengaluru's poor road conditions. Sen's family was also affected when his mother-in-law suffered an injury in a road accident, motivating him to develop a practical solution.

Generative AI also played a significant role in speeding up development. Sen said that building such an application traditionally could have taken months, whereas modern AI tools helped him create a working prototype in around two days.

Could the Pothole Detector Expand Beyond Bengaluru?

The project has already attracted interest outside Karnataka. According to Sen, the Andhra Pradesh government has contacted him to understand the application's reporting system and approach to civic accountability.

Expanding the platform to cities such as Mumbai and Delhi could be the next step. However, scaling it nationwide presents a major challenge: government road-contract data is often fragmented, inconsistent and difficult to access.

Sen currently covers the application's API and server expenses himself and keeps the platform free for users.

If the model can be expanded successfully, Potholes Detector could offer a glimpse of how AI, smartphones and open government data can work together to improve civic accountability. Instead of simply posting pictures of potholes online, commuters could have a technology-driven way to identify the responsible contractor and push for timely repairs.