Turning dozens of raw state revenue records into a property title verdict a banker can defend in a credit committee.

What AdvaRisk does

AdvaRisk builds property and land verification products for Indian banks and NBFCs.

Every home loan, every land backed loan and every project finance deal starts with the same three questions.

<aside>

1. Who owns this property?
2. Is there anything charged against it?
3. Has the title passed cleanly from one owner to the next? </aside>

Banks used to answer those by sending field agents to local registrar offices, photocopying deeds and waiting days for an answer.

The problem

A single order pulls dozens of records across five categories from state revenue databases. Ownership, title chain, mortgage, valuations, other events.

All relevant, none ranked, and no signal about which one decides the outcome. Two records can differ by a middle initial or a transliteration and describe completely different properties.

<aside> 🚨

Problem statement: Operations analysts need to compare near-identical property records fast enough to clear ~20 to 30 orders a day against an SLA, and bankers need a decision they can defend to a credit committee, from the same underlying data.

</aside>

The problem had a shape, not just a description. Miss a charge and a bank lends crores against an encumbered property. Flag something that isn't there and a clean file goes to re-verification, which costs a fee and 2 to 3 days. Those two mistakes aren't the same size, and that asymmetry shaped almost every decision that follows.