A full commercial real estate deal, underwritten end to end
The walkthrough above is a simplified demo of Minervian AI, an AI-powered cash flow modeling tool for commercial real estate. It mirrors the look and flow of the live product across a full deal: lease-by-lease projection, expense recoveries, a multi-tranche capital stack, and an equity waterfall. AI reads your documents; a fixed calculation engine produces every number, so results are reproducible rather than improvised. No login required.
What does Minervian AI do?
Minervian AI builds commercial real estate cash flow models from your documents. It covers office, lab, industrial, and retail assets alongside market-rate multifamily, as single properties or multi-asset portfolios, across core, value-add, and development deals. Its tabs — Buildings, Leases, Market Leasing, OpEx, Lease-Up, Acquisition, Project Costs, Financing, and Exit — produce a monthly cash flow, cash flow by tenant, cash flow by cohort for multifamily, grossed-up OpEx, exit value, levered and unlevered returns, and a 13-section deal summary formatted as investment-memo-ready tables. The walkthrough above is a hard-coded tour of one sample deal; the live product runs on your own.
How long does a first draft take?
Five to ten minutes, from documents to a complete first-draft model, because extraction and autocomplete fill in the assumptions rather than leaving them blank. Building the same model by hand typically takes hours. Prompting a general-purpose LLM from scratch runs 30 to 60 minutes and returns different numbers on each run. Once inputs are in place, recalculation is effectively instant — the engine solves the full monthly projection in a fraction of a second — and you can keep refining any section at your own pace.
Does the AI do the math?
Rather than doing mental math, the AI drives a prebuilt deterministic engine. It reads rent rolls, PDFs, and screenshots and maps what it finds onto defined input fields; the engine then runs the same formulas on every deal, so identical inputs always return identical numbers and each figure traces back to an input you can inspect. General-purpose LLMs re-derive arithmetic on each prompt, which is why their outputs drift between runs. Extraction is a language problem. Underwriting is not.
What financing structures are supported?
Multi-tranche capital stacks including construction, land, bridge, permanent, and mezzanine debt, plus ground leases and seller financing, each with its own sizing and draw schedule. Above the debt sits a configurable equity waterfall built around classes of securities rather than a fixed GP/LP split, so preferred returns, promote tiers, and blended splits can be modeled the way they are actually papered rather than forced into a template.
Who is this built for?
Acquisitions teams at investment, private equity, and private credit firms; capital markets and investment sales brokers; owner-operators, sponsors, and developers; and lenders underwriting the debt rather than the equity. The common requirement is a defensible model fast enough to keep up with deal flow — levered and unlevered IRR, equity multiple, DSCR, and yield on cost, each traceable back to the inputs behind it.
Do I need an account to try it?
No. This page runs the walkthrough with no login, no sign-up, and no credit card. The product itself is web-based with nothing to install, and registration takes a single step, so there is no procurement process or IT ticket between the demo and your own deal. A free account adds your own deals, document upload, Excel export, and saved versions, and costs nothing.