Smt. Sitharaman's District Push and the Property Records Gap

The Hon'ble Finance Minister's argument at the Rajasthan Yuva Conclave was about deployment rather than invention: technologies that work should reach districts quickly instead of stopping at funding rounds and urban pilots. Applied to property-backed lending, that argument runs into a specific obstacle, and it is worth naming precisely.

A borrower in a district town can be creditworthy and the asset sound, and the file will still sit for weeks. Not because the credit assessment is harder, but because the documents that prove ownership live in a state system the lender has no standing connection to, in a language the credit desk may not read quickly.

This is why the same product has two turnaround times depending on the pin code. A metro flat with a clean registered chain and a digitised encumbrance certificate is a days-long check. A peri-urban parcel three districts away, with a mutation entry that never caught up to the last sale, is a different exercise entirely.

What did the Hon'ble Finance Minister call for at the Rajasthan Yuva Conclave?

Smt. Sitharaman called for startup-developed technologies to be taken beyond major cities and deployed in districts and government facilities, arguing that support should not end at funding. Her direct words on the point were general: "Startup ideas, however relevant they may be, need government support, policy support and funding. But more importantly, such ideas should be immediately put to use."

The example the Hon'ble Finance Minister praised in that session was a health deployment, an AI-based early developmental screening tool for children running at 30 primary health centres in Rajasthan. It is worth being accurate about that, because the "every district" line she used was said about taking that screening technology into the public health system, not about lending.

Separately in the same session she highlighted the potential of AI in banking during an interaction with the founder of AdvaRisk. She described technology that helps financial institutions detect potential fraud and monitor loan exposure as having national relevance, and said such solutions should reach banks across the country.

She also encouraged self-help groups to move beyond traditional livelihood activities into businesses such as warehousing and cold storage, supported by bank financing. That is itself a rural lending collateral question, because those borrowers pledge land and buildings rather than receivables.

The through-line is deployment reach. A technology that only works where records are already digitised has not reached a district in any meaningful sense.

Why is a property harder to verify in a district than in a metro?

Because land administration is a state subject, and each state built its own record system on its own timetable, with its own identifiers and extract formats. A lender operating across ten states is dealing with ten systems, and metro properties are simply further along in that digitisation than rural and peri-urban ones.

Three differences do most of the damage. Metro urban property is usually held under a registered sale deed with a traceable chain; district and rural property is often held on revenue records where the mutation entry, not the deed, is the operative document. Metro encumbrance data is more likely to be online for a usable look-back period. And metro files are far more likely to be in English.

None of these is a credit judgement. They are evidence-gathering problems, and they are the reason a good borrower in a small town waits.

Which records does a lender actually need, and where do they live?

A lender needs the ownership chain, the current revenue position, the encumbrance history and the litigation status. Each comes from a different office, and for most states each comes from a different portal.

The operational point is that no single one of these is sufficient. A clean deed with a stale mutation entry is a weak security, because enforcement later depends on the revenue record agreeing with the deed.

What does the Unified Lending Interface change?

It changes the plumbing, not the diligence. The Reserve Bank Innovation Hub, the RBI's innovation arm, exposes a Land Records Service inside its Unified Lending Interface, and in its published API reference that service connects lenders to 11 state government land record databases through a single consistent REST API, so a lender integrates once rather than eleven times. That solves an integration problem. It does not solve a verification problem.

The documented workflow is a submit-poll-read cycle: post a query with state code, district, village and the parcel identifier, receive a request id, then poll until it completes. RBIH documents typical completion at two to eight seconds, a different order of magnitude from a physical record search.

The returned payload defines the boundary. It carries owner details with ownership share, the land parcel's ULPIN and khasra number, area details, and a legal risk block with a mortgage flag and, where present, the name of the bank holding the charge. That is a current snapshot of the revenue record, and it is genuinely useful at the origination screen.

It is not a title investigation. The response contains no chain of prior deeds, no encumbrance history over a look-back period, no litigation position and no lender-format report. A credit officer cannot sanction a mortgage on it, and it was never designed for that. The RBI's own National Strategy for Financial Inclusion frames ULI as digital public infrastructure for frictionless credit, which describes plumbing rather than diligence.

The correct reading is that ULI raises the floor, and the floor is still uneven. The RBI's Trend and Progress of Banking in India report records 64 lenders live on the platform as of 12 December 2025, drawing on digitised land records from eight states, so the single-integration promise applies where a state is connected and nowhere else. Screening a parcel is fast and cheap in those states. Deciding whether the security is good still requires the chain, the encumbrance position and, for most banks, a signed legal opinion.

How do language and script slow a district file down?

Because the record is issued in the state's official language, and the person assessing it usually cannot read that language. This is not a formatting nuisance; it is a translation and interpretation task performed on a legal document, where a misread name or wrong survey number invalidates the check.

The ULI land records response makes the point concretely. Owner details are returned with a name_local field carrying the name in the local script, alongside the father's name in the same script. A Devanagari, Kannada or Bengali string is the authoritative record, and matching it reliably to the KYC name on the application is its own problem, particularly where transliteration is inconsistent across documents.

Multiply that across a title chain of six or seven deeds spanning decades, some handwritten, and the reason district files take longer becomes obvious. The work is legible only to someone who can read the script and knows the state's document conventions.

What does this cost a borrower who is not in a metro?

It costs time and, often, the loan. A file needing a physical search at a distant sub-registrar office, a translated deed chain and a mutation correction can run for weeks, and a proportion of those files are abandoned rather than declined.

The distributional effect is what makes this a policy question rather than an operations one, and it is the same point the Hon'ble Finance Minister was making from the other direction. The borrowers most affected are small businesses and individuals whose main asset is property, in exactly the districts where formal credit is thinnest. Her suggestion that self-help groups move into warehousing and cold storage with bank financing describes borrowers who would pledge exactly this kind of asset.

What can automation not fix?

Automation compresses the gathering step. It does not create records that were never made, and the following limits apply regardless of how good the pipeline is.

• Unregistered instruments. Oral partitions, unregistered agreements to sell and undisclosed tenancies exist outside the registry and outside every API built on it.
• Coverage gaps. ULI's land records service covers a subset of states, and coverage inside a state is not uniform across districts. A parcel outside the covered set still needs a manual search.
• Stale mutation. A revenue record never updated after a sale is a defect in the record itself. Reading it faster does not repair it.
• Chain reconstruction. Assembling a 30-year title chain from prior deeds is interpretive work, and it is the step a title and encumbrance investigation exists to perform. A missing intermediate deed or an unprobated Will is a legal question, not a data question.
• Fresh litigation. A suit filed recently may not appear in any index yet.
• Boundaries. The deed schedule, the recorded extent and the physical boundary can disagree, and only a survey settles it.

Any vendor claiming to have solved district property lending outright is describing the gathering step and calling it the whole job.

Where automated title work fits

Machines gather, normalise, translate and flag, and people interpret. Where a legal opinion is required, an empanelled advocate reviews the evidence and signs it, and that division is what makes automation safe to deploy in the districts where records are messiest.

AdvaRisk states that it operates across 15 states with a database of 250 million properties, and that it builds AI to interpret ownership records across urban and rural areas. It states it was empanelled by the Indian Banks' Association for fraud detection and monitoring large credit, with experience investigating banking exposure of more than ₹1,50,000 crore since 2016. Those are claims about coverage and operating history, and they should be read alongside the limits above rather than instead of them.

For a lending team the practical sequence does not change: screen at onboarding, run automated title and encumbrance work at origination, complete full due diligence before sanction, and keep watching the security afterwards. What changes outside the metros is how much of that can happen without sending someone to a records office. That, rather than any announcement, is what determines whether a district file closes or quietly does not.