ARC Genesis: Gemini Land Intelligence

Gemini turns fragmented land records, planning rules, development potential, and valuation evidence into one auditable property decision.

Comment

Inspiration

Land decisions in India are still made by moving between planning maps, regulations, property records, valuation methods and different professional advisers. We built ARC Genesis because these pieces should work together around one parcel instead of remaining disconnected.

What it does

ARC Genesis is a land-intelligence workspace with four connected parts:

  • Property Search organizes the official-record research workflow while keeping government verification and CAPTCHA as human steps.
  • Potential calculates development feasibility, including zoning, FSI, setbacks, parking, statutory charges and buildable area.
  • Valuation prepares an inspectable professional valuation working draft.
  • Arc AI uses Gemini to understand the request, read the active parcel context, organize facts and guide the user to the correct workflow.

The same parcel remains consistent across these engines. We also created an automated georeferencing workflow for scanned planning maps; 138 plans have been processed.

How we built it

Gemini Flash is used through the Gemini API in our Google Cloud project. It interprets ordinary-language requests, reads project context, structures information and explains available actions.

Regulated calculations are handled by deterministic services rather than being invented by the language model. The application records the provider, model, execution time and a sanitized trace for completed Gemini calls.

The public judge build includes a password-protected demonstration workspace with worked cases covering property search, development potential and valuation.

Challenges we faced

The hardest problem was connecting information that uses different formats and assumptions. Planning maps may be scanned or poorly aligned, regulations contain many conditional rules, and property records often require manual government-portal verification.

We therefore kept sources, assumptions and calculations visible. Gemini coordinates and explains the workflow, while legal verification, CAPTCHA, professional certification and regulated calculations remain bounded.

What we learned

We learned that AI is most useful here when it helps professionals navigate complexity without hiding the basis of a decision. A confident answer is not enough. Users need to see the parcel, source, assumptions, calculation and the point at which human verification is required.

What’s next

Our next step is to run independent professional pilots with architects, valuers, developers and property advisers. We will expand planning-map coverage, strengthen regulatory testing and measure completion time, corrections, repeat use and willingness to pay.

The product is still being validated commercially. The INR 500 received during the competition period was a retained team-insider payment-gateway test, not an independent customer sale.

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for ARC Genesis: Gemini Land Intelligence

Built With

Share this project:

Updates