Funding has followed the products. Indian AI start-ups raised more than $1 billion in the first nine months of this year, the BBC reported, even as fundraising for the wider start-up sector stayed relatively muted. Investment in AI start-ups rose 265% from a year earlier, according to Inc42, a start-up information platform.
Over the following months he moved between government offices to obtain about a dozen records that would show the property belonged to his family. Frustrated by the search, he built Landeed, an artificial intelligence platform that aggregates fragmented land records from many sources, verifies titles and runs legal checks across 24 Indian states.
The BBC reported that Landeed now holds 1.8 billion documents, the largest repository of title data in the country, and is used by millions of individual and enterprise customers in real estate transactions.
The gaps matter. Property disputes account for a large share of civil litigation in India, where some 80% of household wealth is held in real estate, and the records system is open to misuse.
"A sale deed merely records the transaction - multiple other documents may be needed to prove title,"Mandava told the BBC. India follows presumptive titling, under which the buyer, not the state, must establish ownership.
Landeed has mapped large parts of the country with geospatial technology and data from state land revenue departments, registration offices, survey records and municipal databases. It links registered properties to official land identification numbers, survey numbers and ownership records.
The company is one of a growing group of Indian start-ups using AI on problems in legal technology, health care, infrastructure, education and agriculture. Dozens are working with fragmented, multilingual data in industries where costs are tightly constrained.
Road damage is another target. A tool that detects potholes, identifies the civic officials responsible for repairs and sends them the data recently drew viral attention, a reflection of how widespread the problem is in Indian cities. At least half a dozen similar services now map and report road damage to authorities.
Other start-ups are translating tribal languages and assisting overworked public health staff with medical tasks. One health app produces simplified, AI-generated summaries of lab reports in several languages, helps users find nearby facilities matched to their needs and gives doctors pre-consultation support, including medical history and medication details.
Through ties with municipal hospitals, that service also gives administrators a public health dashboard built from aggregated, anonymous lab data, offering visibility into emerging infections and other disease trends.
Computing power has not kept up. India has promised start-ups 38,000 graphics processing units through its government-led AI mission, but demand for that capacity is far greater than supply, the BBC reported. High-quality chips at the right scale and price remain a major barrier for many founders.
Experts warned that if those constraints are not addressed, Indian AI companies risk building shallow solutions that a next generation of founders will quickly make redundant.