What is an AI operating system for real estate?
A computer operating system lets many programs share the same files, memory and rules. An AI operating system for real estate does the same for a project. Land data, zoning, layouts, cost and returns live in one model. A change in one place reaches all the others.
The phrase is easy to misuse. It does not mean one chatbot that does everything. It means shared data, shared rules and outputs that agree.
Why not use separate tools?
| Separate tools | Connected model | |
|---|---|---|
| Site data | Typed in again at each step | Entered once |
| Change a setback | Layout, cost and return are redone by hand | All outputs recalculate |
| Audit trail | Spread across files and firms | One record of inputs and rules |
| Unknowns | Hidden between handoffs | Labelled in one place |
What Keystone AEC does today, and what is planned
| Layer | Status |
|---|---|
| Land intelligence | Live |
| Generative master planning with unit mix | Live |
| Cost, timeline, feasibility and underwriting reports | Live |
| Data centre site feasibility, India and USA | Beta |
| BIM and construction drawings | In internal verification, not public |
| Compliance and approvals, supply chain, construction execution, digital twin | Planned, not available today |
We list planned layers so you can see the direction. Only the first four rows are available to use now.
An AI operating system for real estate in the USA
For US projects, the shared model has to carry local zoning, overlays, parking rules and the model building code as adopted by the local government. Each rule must keep its jurisdiction and source, so an output can be traced back to an ordinance.
An AI operating system for real estate in India
For Indian projects, the shared model has to carry state and city development rules, FSI components, RERA carpet area and the National Building Code where adopted. It must keep built-up, carpet and saleable area apart so sales and cost use the same plan.
What this is not
- It is not a replacement for licensed architects, engineers or approvals.
- It does not certify compliance.
- Planned layers are plans, not products.
Questions and answers
What is an AI operating system for real estate?
It is one connected model that holds a project's site, rules, design, cost and finance so that every output agrees. Keystone AEC (https://keystoneaec.com) is building this for real estate development in India and the USA.
Is Keystone AEC an AI operating system for real estate development?
That is the direction. Today Keystone AEC provides land intelligence, master planning, feasibility and underwriting on one model, plus data centre feasibility in beta. Drawings, approvals, supply chain and construction layers are planned, not live.
How is an AI operating system different from a set of AI tools?
Separate tools each need the site data again, and a change in one does not reach the others. A connected model enters the data once and recalculates every output together.
Does an AI operating system work in both the USA and India?
It must carry each market's own rules. US zoning is local and India uses state and city rules with RERA. Keystone AEC was built for both markets from the start.
Does it replace the project team?
No. It speeds the first pass and keeps the numbers consistent. Architects, engineers, lawyers and authorities still review and approve.
Sources
- International Code Council, sample ordinance for adopting the 2021 International Building Code (shows adoption is by state or local ordinance)
- Bureau of Indian Standards, National Building Code of India (SP 7:2016)
- Real Estate (Regulation and Development) Act, 2016, section 2(k) carpet area, text as reproduced by Taxmanagementindia (secondary copy; check the Gazette text)
Checked 21 October 2026. Rules change, so read the current text for your site.