End-to-end AI underwriting for institutional commercial real estate
Bring a prompt, rent roll, offering memorandum (OM), PDF, or screenshot. The AI reads your documents and proposes the assumptions they leave out, while our proprietary deterministic engine delivers the complete cash flow projection, returns, and deal summary.
See how it works →What Minervian AI does
Minervian AI is end-to-end underwriting software for commercial real estate, built for institutional deals. It carries a deal from the offering memorandum through extraction, proposed assumptions, and review, to a complete monthly cash flow model with debt and an equity waterfall, and out to Excel as investment-memo-ready tables. Office, lab, industrial, and retail are covered alongside market-rate multifamily, with several buildings and several asset classes inside one deal, on core, value-add, and development deals.
Modelling depth at a glance
| Leases | Projected individually — base-year stops, expense stops, escalations, free rent, TI, leasing commissions, renewal probability, downtime |
| Recoveries | Grossed up per tenant off a full operating expense budget, not a flat percentage |
| Debt | Multi-tranche: construction, land, bridge, permanent, mezzanine, ground lease, seller financing. Fixed or floating on a forward SOFR curve, with caps and swaps |
| Equity | A waterfall with no fixed GP/LP template — any number of classes, tiers, and hurdles |
| Multifamily | Cohort modelling for value-add at unit-type level, with per-band renovation cost, downtime, and post-renovation rent |
| Development | Hard costs, soft costs, land, and contingency on an S-curve, with construction interest and lease-up in the monthly cash flow |
| Outputs | Monthly cash flow, by tenant, by cohort, grossed-up OpEx, exit value, and investment-memo-ready tables: capitalization and sources and uses, yield on cost, profits and returns, waterfall, budget, revenue and operating expenses, and financing |
An AI assistant that builds the model with you
Describe the deal the way you would explain it to a colleague, from a prompt, rent roll, OM, PDF, or screenshot. The AI does considerably more than read those documents:
- Proposes every missing assumption across the whole model.
- Edits inputs from plain-language instructions, including batch changes across many leases or loans at once.
- Flags gaps and contradictions before they reach the output.
- Explains how any section calculates, on request.
- Draws on the underlying model's own knowledge of market norms to propose a defensible figure where you do not have one.
What it never does is the arithmetic. A proprietary deterministic engine performs every calculation, so identical inputs return identical numbers and every figure traces back to an input you can inspect.
The waterfall is the clearest case. When a managing director says "15 percent preferred to the LP, GP catches up, then 18 percent, then the GP's special bonus, then 25 percent, then all residuals to the GP," that sentence goes straight into the chat and the structure exists seconds later. Building it by hand is two to three hours of an excellent associate's time, and longer still if the template will not bend. Re-timing a draw schedule or adjusting a whole set of leases works the same way.
Autocomplete fills what the OM leaves out
A typical OM supplies only 30 to 60 percent of what a complete underwriting needs, so Autocomplete proposes the rest across the whole model — leasing, operating expenses, lease-up, financing, fees, exit, and the waterfall — marking each value as proposed so it is obvious which numbers came from your documents and which were filled in. That is where most of the time saving comes from.
Portfolios and development
Portfolios are where that matters most. A deal holds as many buildings as it needs, and changing an assumption across ten properties is one instruction to the assistant, applied in seconds — the same edit across ten separate model files is a twenty-five-minute job of exporting, pasting, and repairing rows that no longer line up. Development deals run a full monthly cash flow too, with S-curve hard cost draws, construction interest, and lease-up, rather than a simplified budget bolted onto a stabilised model.
How it differs from the alternatives
Established CRE modelling software gets the depth right, but assumptions are keyed in by hand over hours, it runs on the desktop, and it is commonly licensed in the thousands of dollars per seat per year. Prompting a general-purpose LLM starts faster, but the numbers move between runs and it cannot hold grossed-up recoveries, a multi-tranche levered cash flow, or a promote waterfall — a spreadsheet it writes will not carry them either. Minervian AI uses a language model for the language work and a fixed engine for the arithmetic, which is what lets a five-to-ten-minute first draft also be an institutional one, at $25 per seat per month with every rate published.
Full capability detail, tool comparisons, and FAQ on the platform page →
End-to-End Underwriting
From offering memorandum to investment-memo-ready tables in one system: extraction that maps deal terms onto the right input assumptions, proposed values for whatever the documents omit, validation that flags gaps and contradictions, a complete monthly model with debt and an equity waterfall, and Excel export. No handoff to a spreadsheet in the middle, and no separate modelling tool at the end.
Built for Real Complexity
Hundreds of rows and columns for true cash flow analysis, including lease-by-lease projection with probability-weighted renewals, expense recoveries grossed up per tenant, and leasing costs by tenant. Multiple buildings and multiple asset classes inside a single deal, across office, lab, industrial, retail, and market-rate multifamily, on core, value-add, and development deals.
Reliable, Deterministic Math
No black boxes, no invented logic. The AI reads and structures your deal, but it never does the arithmetic: every number flows from your inputs through the same fixed formulas on every deal, producing the same result every time without hallucination.
Extraction in Minutes, Calculations in Seconds
A complete first draft takes five to ten minutes, and once the inputs are in place the model runs in a split second, producing full cash flow projections, IRR, equity multiple, DSCR, and deal summaries instantly, not hours later.
Checkable, Not Just Trusted
Reproducibility is the floor: identical inputs return identical numbers, so a figure that moves means an input moved. Above that, the intermediate schedules are outputs in their own right rather than something hidden inside a result — the grossed-up operating expense budget behind every recovery, cash flow by tenant and by cohort, and a per-tier waterfall schedule showing opening balance, accrual, and what each tier was paid, which foots line by line. Every figure traces back to an input you can change, and the assistant explains how any section calculates on request.
No Context Bloat
Proprietary data schema ensures your context is not bloated, keeping token usage down, latency minimal, while maintaining model coherence.
Excel Compatible
Handles the full underwriting end-to-end, including the parts that are not practical to maintain in a spreadsheet at all: lease-by-lease projections across hundreds of tenants, recoveries grossed up per tenant, and leasing costs resolved per lease. Bring documents in as Excel, CSV, PDF, or a screenshot; export the cash flow projection, deal summary, output tabs, multifamily rent roll, and equity waterfall to Excel to drop into your existing workflow.
Data Security
Industry-standard encryption, hardware-backed security keys on Google Cloud, and multi-factor authentication. Your data is never used to train AI models.
See the full flow live.
Best experienced in desktop or horizontal format.
Watch the AI extract, confirm, and model a deal from a single description.
Intelligent Ingestion
PDFs, screenshots, natural language: feed the deal however it comes to you. The AI reads it, structures it, and populates your model inputs automatically.
Institutional-Grade Analysis
Proprietary engine runs your full underwriting the moment data is ready: lease projections, expense recoveries, debt modelling, IRR, multiples, and sources & uses, all in seconds.
Outputs Built for Decisions
Deal summary with returns, budget, profits, and sources & uses. Detailed monthly cash flows broken out by individual tenant.
Get Started with Minervian AI
Log In
Register
Future Features
New Asset Classes
Live today: office, lab, industrial, retail, and market-rate multifamily. Upcoming: condo developments, hotel, student housing, senior living, self-storage.
Asset Management
Track and manage assets post-close, keeping your underwriting and operational data in one place.
Portfolio Management
Aggregate across deals, analyze fund-level performance, and synthesize insights across your entire portfolio.
External Sharing
Collaborate beyond your team by sharing deals securely with lenders, partners, and investors; no account required on their end.
Submit via Email
Email your inputs to the platform and get your analysis back automatically. No login, no browser, just results in your inbox.