JupiteX Get the app
Science & Technology17 Nov 2026 · about 7 min

The AI start-ups fixing India's broken roads and messy land records

The brief

India’s AI start-ups are using technology to tackle problems that affect daily life and public services. Their work often focuses on consumer applications built from fragmented, multilingual data, rather than research into frontier language models. The goal is to make scattered information useful for decisions and action. Landeed, for example, combines geo-spatial mapping with records from land departments, registration offices, surveys and municipal databases. It links properties to official identifiers and ownership records, then uses AI to read titles and flag inconsistencies. In Bengaluru, Gaurav Sen’s app uses dash-cam images, satellite navigation and an AI vision model to detect potholes and identify the officials responsible for repairs. Other applications include Fractal Analytics’ Vaidya, which supports patients, doctors and hospital administrators in Mumbai. AI start-up funding rose 265% year-on-year, but firms still face difficulty turning useful tools into commercial services beyond pilot projects. Their reach will depend partly on whether customers can afford them.

01

How are Indian AI start-ups using technology to tackle problems such as potholes and fragmented land records?

India’s AI start-ups are using technology to tackle problems that affect daily life and public services. Their work often focuses on consumer applications built from fragmented, multilingual data, rather than research into frontier language models. The goal is to make scattered information useful for decisions and action.

Landeed, for example, combines geo-spatial mapping with records from land departments, registration offices, surveys and municipal databases. It links properties to official identifiers and ownership records, then uses AI to read titles and flag inconsistencies. In Bengaluru, Gaurav Sen’s app uses dash-cam images, satellite navigation and an AI vision model to detect potholes and identify the officials responsible for repairs.

Other applications include Fractal Analytics’ Vaidya, which supports patients, doctors and hospital administrators in Mumbai. AI start-up funding rose 265% year-on-year, but firms still face difficulty turning useful tools into commercial services beyond pilot projects. Their reach will depend partly on whether customers can afford them.

02

How does Landeed combine maps, official records and AI to check property documents?

Landeed is a digital platform designed to bring scattered land information together. This matters in India because a sale deed alone may not establish ownership; buyers can be asked for several other records. Landeed aims to make it easier to check property titles and carry out legal checks.

The platform maps large areas using geo-spatial technology and data from state land revenue departments, registration offices, survey records and municipal databases. It links registered properties to official land identification and survey numbers, as well as ownership records. An AI layer then reads title documents, extracts information and looks for inconsistencies. The article says this analysis can take minutes, rather than the several days lawyers and brokers may need.

Landeed operates across 24 Indian states and has a repository of 1.8 billion documents. Millions of individual and enterprise customers use the app for real estate transactions. Its approach depends on bringing records from different sources together; the article does not claim that AI itself guarantees a title or resolves a dispute.

03

How large is Landeed’s title-document collection, and across how many Indian states does it operate?

Landeed has built a repository of 1.8 billion title documents, according to the article. Its app operates across 24 states in India. That scale is notable because property information is spread across different government and municipal sources, rather than held in one easily checked record.

The platform uses geo-spatial technology to map properties and gathers data from state land revenue departments, registration offices, survey records and municipal databases. It links registered properties to official land identification and survey numbers and to ownership records. An AI layer can then read titles, extract information and spot inconsistencies within minutes. The article contrasts this with several days for lawyers and brokers to do the work.

Millions of individual and enterprise customers use Landeed while carrying out real estate transactions. Its size reflects demand for tools that can aggregate and examine fragmented records. The article gives the current scale, but does not specify how many documents are added over time or how many customers use the service in each state.

04

Why can a sale deed alone be insufficient to prove ownership of land in India?

A sale deed can be insufficient because India follows presumptive titling, as described by Landeed founder Sanjay Mandava. Under this system, the buyer is responsible for establishing that the property belongs to them. The sale deed records the transaction, but it does not necessarily provide all the evidence authorities may request.

Mandava experienced this firsthand after travelling from Chicago to Chennai to resolve a family land dispute. He had to visit government offices to obtain a dozen records. Authorities asked for documents beyond the deed, including building certificates, property tax receipts and land survey maps. Landeed was built in response to the difficulty of gathering and checking this fragmented information.

The article says property disputes account for a bulk of civil litigation cases in India, where about 80% of household wealth is held in real estate. Landeed aggregates records from multiple sources and uses AI to check titles for inconsistencies. The tool can speed up document review, but the article does not say that it replaces the buyer’s responsibility to establish ownership.

05

What can happen when AI tools detect potholes and send their locations to the officials responsible for road repairs?

AI pothole tools can help turn road damage into information that reaches the officials responsible for fixing it. The article describes a system that detects potholes, identifies the civic officials responsible and sends them the data. That can make road conditions more visible to those who may act on the reports.

Gaurav Sen’s app uses a dash-cam, satellite-based navigation and an AI vision model to analyse images. It detects potholes and identifies the relevant officials. Sen told the BBC that a municipal corporation could use the app on public buses and taxis to access real-time, updated road-condition data. Several other services also map and report road damage to authorities across India.

The article notes that potholes are widespread in urban areas and that Sen’s app received viral attention. Sending locations and road-condition data can support municipal monitoring, but it does not itself repair a road. The article describes reporting and possible use by civic bodies; it does not establish how quickly officials respond or how many reported potholes have been fixed.

06

What makes it difficult for Indian AI start-ups to grow beyond pilots, even when their tools are useful?

A central challenge for Indian AI start-ups is turning a promising pilot into a service that customers will pay for at scale. Fractal Analytics co-founder Srikanth Velamakanni told the BBC that building in India can cost more than Indian customers are willing to pay. That gap makes commercialising useful ideas difficult.

One pressure is the cost of computing. Velamakanni said the computing power needed to run an AI model is the same whether it is built in India or the US. But he said American customers have a far bigger paying capacity. The article also describes start-ups working in fragmented, multilingual datasets and cost-sensitive industries, where solutions must serve local needs while remaining affordable.

Funding for AI start-ups rose 265% year-on-year, and they raised more than $1bn in the first nine months of the year. That growth has not removed the commercial hurdle: the article says firms face an uphill task taking ideas beyond pilots. It ends by noting that founders are being forced to build around scarcity, but the provided text does not complete that point.

07

What is presumptive titling, and how does it differ from a system where the government guarantees a property’s ownership?

Presumptive titling is the system Mandava describes for India: the buyer is responsible for establishing ownership of a property. A sale deed records that a transaction took place, but it may not be enough to prove title. Buyers can need several other documents to demonstrate that the land is theirs.

Mandava contrasted this with the US and UK, where he said the government verifies ownership when a property is purchased. In India, authorities may ask for a range of supporting records, including building certificates, property tax receipts and land survey maps. When Mandava travelled to Chennai to handle a family land dispute, he spent months visiting government offices to access a dozen records.

This distinction helps explain why checking property documents can be difficult and why Landeed was created. It aggregates records from government and municipal sources, links properties to official identifiers and uses AI to spot inconsistencies in titles. The article says the app can speed up checks, but it does not describe a change to India’s titling system or say that the platform guarantees ownership.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web