Valuation on famously unreliable data.

Indian property data is thin, inconsistent and frequently understated for tax reasons. Any valuation model that does not treat that as its central problem is fitting a curve to fiction.

01 / What is hard here

The problem underneath the problem.

Recorded transaction values often differ materially from actual consideration, listings persist long after a property has sold, the same building is described five different ways, and comparable-sale density is low outside a handful of micro-markets.

The consequence is that a point-estimate valuation is misleading regardless of how good the model is. The defensible output is a range with an explicit confidence, and an honest refusal to value where comparable evidence does not exist.

02 / What we build

Work we take on in real estate.

Valuation models: Automated valuation with prediction intervals and abstention in thin markets, calibrated per micro-market rather than nationally.
Listing data quality: Deduplication, entity resolution and attribute normalisation across sources — which is the bulk of the work and the precondition for everything else.
Lead scoring: Ranking enquiries by likelihood of genuine intent so that limited sales attention is spent where it converts.
Portfolio and rental analytics: Yield, occupancy and rent-roll analysis for institutional owners and managers.
Document processing: Extraction from title documents, agreements and approvals, with human review on anything that carries legal consequence.

03 / Constraints

What shapes the build.

These are design constraints agreed at the start, not compliance work bolted on before launch. Retrofitting them is expensive and usually incomplete.

Data provenance: Registry, listing and broker sources disagree systematically. Which source is authoritative for which field is decided explicitly and documented.
Thin markets: Where comparable evidence is insufficient, the correct output is a refusal to estimate, not a wide guess presented as a number.
Fair-housing exposure: Location features correlate strongly with community composition. Models touching access to housing or finance are reviewed for that before deployment.
Valuation liability: An automated valuation informs a professional judgement. It does not replace a formal valuation where one is legally required.

04 / What we advise against

Where we would say no.

We would advise against publishing point valuations without intervals, and against any lead or tenant scoring that uses location as a proxy for community characteristics.

We would rather lose the work than build something that creates a liability for you or a harm for the people on the other side of it. If that rules out what you had in mind, it is better established now than at delivery.

05 / Other sectors

Adjacent work.

Different domains, the same delivery discipline — a measured baseline, a real evaluation set, and monitoring that catches decay early.

Next step

Working in real estate?

Tell us which process is slow, expensive or error-prone. The diagnostic will tell you whether this is a problem worth solving with machine learning — including if it is not.

Request a diagnostic All industries