Built for the third shift.
Paper gave way to Excel. Excel gave way to the ERP. Built AI exists because the third shift, from the ERP to an AI-native operating system, is starting now, and someone has to build the platform it lands on.
Institutional real estate runs on systems that were never built to talk to each other. The models sit in one place, operations in another, and everything in between lives in spreadsheets, so the truth of a portfolio lives in none of them. Every firm pays for this twice: once in the analysts who re-key it, and again in the decisions made on stale numbers.
We build the layer that sits on top: one knowledge graph across a firm's systems and documents, a deterministic engine that computes on it, and a catalog of agents that do the work, underwriting, abstraction, variance, reporting, with every number cited to source and every external action approved by a human.
We founded the company in 2020 and work with institutional owners, investors, and managers across the US, UK, and Europe, from New York and London. The full argument is in the AI-native operating system thesis.
Natan Lempert
More than fifteen years across the real estate buy side and in investment banking. Natan began at Goldman Sachs and Citi in London, advising on over $20B of M&A, debt, and equity transactions, then worked value-add and opportunistic deals across Europe at Benson Elliot, and most recently ran both acquisitions and asset management at an investment manager backed by Korean capital, sourcing and executing over €2.5B of deals.
Connect on LinkedIn
Damien Georges
Twenty years building the data infrastructure of institutional real estate. Damien founded three companies in the space, Voyanta (acquired by Altus), Hipercept, and FUEL (both acquired by RealPage), then ran asset and investment management solutions as an SVP at RealPage. At Built AI he leads the US business from New York.
Connect on LinkedInEvery number cited.
Output that cannot be traced to a source document does not ship. Provenance is the product, not a feature of it.
Nothing sends without a human.
Agents draft, reconcile, and flag. A person approves anything that leaves the building. That line does not move.
Deterministic, not probabilistic.
The math runs on an engine, not on a guess. Same inputs, same outputs, every time, at audit depth.
Sit on top, not rip out.
The systems a firm already runs stay in place. We unify them into one source of truth instead of asking anyone to rip anything out.
In December 2025 we raised a $6M seed round led by Work-Bench, with Lerer Hippeau and Emerald Pine, to scale the platform across the US and UK.
Coverage of the round, our awards, and the media kit live on the press page.
See it on a portfolio shaped like yours.
We'd love to walk your team through a live Built AI environment, the knowledge graph, the engine, and the agents, on assets that look like your own.