A Bengaluru software entrepreneur has built an app that can detect potholes through a phone or dashcam, locate them, identify the road contractor using public tender data and prepare a complaint to the Greater Bengaluru Authority. The experiment offers a useful view of the city’s road crisis: technology can make reporting easier, but it cannot by itself ensure that a complaint is accepted, assigned, acted upon and closed only after the road is repaired.
Gaurav Sen, a 33-year-old entrepreneur who runs AI Engg and InterviewReady, began developing Pothole Reporter after repeatedly encountering damaged roads on his drives between Cunningham Road, HSR Layout and Hebbal. On one routine journey, he counted 12 potholes. He told The Indian Express that Bengaluru’s roads had become a serious problem, particularly for two-wheeler users navigating between heavy vehicles, dividers and uneven surfaces.
The app is a small, approximately 10 MB application that can use a phone mounted on a dashboard or connected to a dashcam. Sen said it can process a video stream, including footage from Meta Glasses, and identify a pothole in an image. It then geolocates the defect and uses public Karnataka government data on road tenders to determine which contractor is responsible for the road. The details, along with images, are compiled into an email addressed to the Greater Bengaluru Authority commissioner.
That sequence is significant because a pothole report is not simply a matter of identifying a hole in the road. A functioning complaint needs a location, evidence, an administrative recipient and, ideally, a responsible agency or contractor. By attempting to combine these elements, Pothole Reporter is designed to reduce the gap between what a citizen sees on the street and what a public authority needs to process a complaint.
The app also reflects a wider shift in how civic problems are being documented. Smartphone cameras, location services and artificial-intelligence tools allow residents to collect evidence at a scale that municipal field inspections may not easily match. Sen said the initial version was created in a few hours with help from OpenAI’s Codex, followed by several days of manual review and changes to the generated code. He also said the application has been designed to reduce the risk of incorrect AI outputs, commonly called hallucinations.
But detection is only the first stage of a road-maintenance system. The more difficult question is what happens after a complaint is sent. Sen said an email he sent in mid-August had not been resolved when he spoke to The Indian Express. He is concerned that without application programming interface access to the authority’s system, Pothole Reporter may simply generate emails that are not integrated into the official workflow.
This distinction matters for Bengaluru’s civic technology landscape. A platform can improve the quality of information entering government without improving the government’s ability to respond. If a report arrives outside the authority’s case-management system, it may not receive a tracking number, an assigned officer, a repair deadline or a closure record. The citizen may have supplied better evidence, but the administrative chain remains unchanged.
Sen’s criticism of existing reporting tools points to the user-experience side of the problem. He said the BBMP Sahaya app requires several details and can be difficult to use. A more user-friendly version is planned, he said, but is not yet a reality. The difficulty is not trivial. Reporting a pothole requires time, attention and sometimes physical risk, especially when a person must stop near moving traffic to photograph a damaged stretch. If the process is cumbersome, many residents will choose not to report defects at all.
Even when residents do report them, unresolved complaints can weaken participation. Sen said people become exhausted when they repeatedly photograph potholes and receive no visible result. He also raised the possibility that complaints could be marked closed on a platform without the pothole actually being fixed. That concern shifts the focus from access to accountability: the measure of a civic app is not the number of complaints it receives, but whether the system can demonstrate what happened to each one.
The road-safety consequences make this more than a question of digital convenience. The Indian Express report referred to the death of a woman in Ananth Nagar near Hebbagodi after the two-wheeler on which she was travelling as a pillion rider skidded and a mini goods vehicle ran over her. Residents who protested after the incident pointed to poor road conditions and the lack of safe alternatives. The report does not establish that a pothole caused the death, but it places the complaint about road quality within a wider public-safety concern.
Sen’s own experience also shaped the project. He had previously navigated potholes in Mumbai and said his mother-in-law suffered a severe knee injury after a two-wheeler slipped on an oil-slick road. His account illustrates why road defects are experienced differently by different users. A pothole can be an inconvenience for a car driver, a loss of control for a two-wheeler rider and a serious hazard when traffic, road edges and dividers leave little room for avoidance.
The use of public tender data introduces another institutional question. Road maintenance responsibility is often difficult for residents to identify because roads may involve different authorities, contracts, construction periods and maintenance obligations. Pothole Reporter’s attempt to connect a road location to a contractor is therefore an effort to make responsibility visible. However, the source material does not establish how complete or current the tender data is, how responsibility is resolved when multiple agencies overlap, or whether a contractor can be compelled to act through the generated complaint.
That uncertainty is central to the app’s potential. Technology can identify a defect, but the authority must still verify it, determine jurisdiction, establish liability, issue instructions and inspect the completed work. Each step can fail independently. A location may be inaccurate; a tender record may not correspond to the current road condition; a complaint may reach the wrong office; or a repair may be temporary and deteriorate again. The more links a system contains, the more important it becomes to record status at every stage.
The problem also exposes the limits of treating civic applications as isolated products. Sen said senior figures from several municipalities and some politicians had contacted him, indicating interest in the idea. Yet he added that action requires more than interest. For a reporting tool to become part of municipal administration, it would need an agreed intake channel, technical integration, clear ownership, response standards and a way for citizens to verify outcomes. None of those requirements is solved simply by making the application easier to download.
Bengaluru’s road problem is therefore also a data and governance problem. Residents can see potholes directly, but the public system needs structured information: where the defect is, which road it affects, who maintains that road, when it was reported, what action was ordered and whether the repair survived inspection. A civic platform that captures only the first part of this chain may improve visibility without creating accountability. A platform connected to the full chain could make maintenance performance easier to measure.
The source report does not provide a citywide count of potholes, a response-time comparison between government platforms or evidence showing whether Pothole Reporter has led to completed repairs. The app was still being prepared for approval on the Android app store and was available on GitHub at the time of publication. Its effectiveness, therefore, remains an open question rather than an established outcome.
What the project already demonstrates is the gap between citizen initiative and institutional capacity. A resident can now use artificial intelligence to detect road defects, combine images with geographic information and identify a possible contractor in a short time. The unresolved challenge is whether the public authority can receive that information in a system that assigns responsibility and records a verifiable repair.
For Bengaluru, the next test is not whether more potholes can be found. It is whether every credible report can move through a transparent administrative process and end with a repair that citizens can confirm. Until that connection exists, civic technology may make the city’s road failures more visible without making the roads safer.
