# Minervian AI > End-to-end AI underwriting software for commercial real estate, built for > institutional deals. Carries a deal package from extraction through proposed > assumptions and review to a full monthly model with debt and equity waterfall, > and out to Excel as investment-memo-ready tables. Deep lease-by-lease, recovery, > financing, and waterfall modeling; a complete first draft takes 5-10 minutes. Minervian AI (legal name: Minervian AI Inc.) is a web application. The AI layer does considerably more than extraction: it reads documents, proposes the assumptions those documents omit, adds and edits inputs from plain-language instructions, validates the model for gaps and contradictions, explains how any section calculates, and applies market norms drawn from its own training rather than a live web search. What it does not do is the arithmetic. A prebuilt deterministic engine performs every calculation, so identical inputs always return identical numbers and each figure traces back to an input the user can inspect. The AI cannot invent line items or formulas. ## Key facts - Scope: end-to-end. Deal package -> extraction onto structured inputs -> autocomplete for missing assumptions -> validation and review -> full monthly model with debt and equity waterfall -> Excel export as investment-memo-ready tables. No handoff to a spreadsheet in the middle, and no separate modeling tool at the end. - Built for institutional deals. A 2,000,000 SF office tower with 150 tenants, mixed lease structures, base-year stops, escalations, renewal options, free rent, TI, leasing commissions, renewal probabilities, downtime, and market rent varying by floor and space type is a supported case, not an edge case. Sold self-serve rather than through an enterprise sales cycle; that is a go-to-market choice, not a limit on modeling depth. - Breadth: five asset classes, three deal types, and single assets or multi-asset mixed-use portfolios, all in one model. - Time to first-draft model: 5-10 minutes. Building the same model by hand typically takes hours. - Asset classes: office, lab, industrial, retail, and market-rate multifamily. - Deal types: core, value-add, and development. - Portfolios: multiple buildings per deal, each with its own asset class; mixed-use modeled in a single projection. - Debt: multi-tranche capital stacks including construction, land, bridge (commercial and MF), permanent, and mezzanine, plus ground leases and seller financing. Bridge facilities support separate funding buckets per use and different funding mechanisms (equity-first, pari passu). - Loan sizing bases: fixed amount, % of purchase price, % of total project cost, LTV against stabilized value, debt yield target, DSCR target, or sized off the balance of one or more prior loans. Tranches added in one click and reordered by dragging to change seniority. Accordion / re-draw supported. Interest rate caps and swaps supported. - Financing fees, each as a percentage or a fixed amount: origination, exit, broker, guarantee, legal, third-party reports, defeasance, taxes, extension, and the upfront interest rate cap premium. Carried through sources and uses and through levered cash flow rather than collapsed into one cost of debt. - Other income: modeled per stream because it commonly drives around 5% of property value. Unit-level streams (RUBS, cable) carried per unit; project-level streams (antenna, signage) stand alone. Parking held as inventory by type, occupied spots earning from day one and vacant spots absorbing on their own schedule, transient and long-term treated differently. Parking committed inside a commercial lease is tagged and reconciled against the deal-level parking pool so no stall is double-counted. - Equity: waterfall with no fixed GP/LP template. Any number of classes of securities, investors, and tiers. Can be described in plain language rather than assembled tier by tier. - Development: modeled as a full monthly cash flow, not a budget bolted onto a stabilized model — construction interest, draws, and lease-up run through the same monthly projection. Hard costs, soft costs, land, and contingency, with S-curve spend modeling for hard costs and flexible draw timing. - Commercial leases: percentage rent, stepped rent, and explicit rent schedules; gross, net, fixed, expense stop, and custom recovery structures; all leasing commission structures including New York style tiers; tenant improvement allowances. Leases are projected individually. - Recoveries: calculated from a grossed-up operating expense budget rather than a flat percentage, resolved per tenant, with the grossed-up schedule exposed as its own output. - Property cash flow rigor: follows stringent and comprehensive standards of institutional practice, distilled from years of underwriting inside institutional shops (global development, large-scale urban development, and private equity managing institutional LP capital and JV waterfall structures). Monthly and lease-by-lease; renewal and new-lease scenarios probability-weighted rather than blended; TI, leasing commissions, free rent, and downtime resolved per lease and per renewal; general vacancy, credit loss, and cost inflation are explicit inputs rather than one haircut at the end. - Not a screening tool. This is a full institutional underwriting, not a go/no-go signal and not a generalized pro forma. Output is a complete monthly model presented as a 13-section deal summary and exported to a workbook running to dozens of tabs (around twenty on the inputs side alone). - Multifamily: rent roll editable unit by unit; a unit summary giving control at unit-type level; market leasing assumptions per unit type; lease-up pace per unit type or global; cohort modeling for value-add. - Multifamily value-add: bands specified as ranges below market, with units bucketed into cohorts automatically. Each band carries its own turnover pace, time offline for capital work, capex per unit, cash-for-keys budget, lease-up free rent, and leasing commissions. Tracked at unit-type level on a month-by-month take-down schedule across a multi-year rollout, not at building level. Vacant and rent-restricted units handled separately. Parking and other income modeled per stream. - Deliberately NOT projected unit by unit. A per-unit projection implies knowing which apartment turns in which month; business plans are written as an intention to turn units 20-40% below market at some monthly pace, so the model takes the same shape. More granular than a building-level projection, more honest than a unit-level one. - Exit: adjustable for post-sale capital expenditure and leasing costs; each saleable entity carries its own cap rate and timing, so buildings can sell on different dates. - Loan mechanics: fixed or floating rates, floating priced off a forward SOFR curve (1M or 3M) with credit spread and rate caps; interest-only, amortizing, and interest-only then amortizing; PIK / accrual interest; refinancing that retires one or several prior loans and is sized off the outstanding balance; extension fees; partial discharges on sale. - Waterfall mechanics: capital contributions by class, preferred return accruing monthly, return of capital, GP catch-up, any number of promote tiers struck on deal-level IRR, equity multiple, profit share, or a fixed dollar amount, refinancing proceeds and sale proceeds distributed through the same tiers, and a residual split taking everything remaining. - Lease mechanics: base-year stops, expense stops by PSF or total, capped OpEx growth over base year, annual percentage / $ per SF / inflation-indexed / explicit rent schedules, free rent, downtime, tenant improvements, renewal probability, renewal options and nested extensions, and market rent by floor or space type via multiple market leasing sets. - Checking a number: identical inputs return identical numbers, so a figure that moves means an input moved. Intermediate schedules are outputs in their own right, not 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, footing line by line. Every figure traces back to an editable input, and the assistant explains how any section calculates. - Portfolios: changing an assumption across ten properties is one instruction applied in seconds; the same edit across ten separate model files is roughly a 25-minute job. - Return metrics: levered and unlevered IRR, equity multiple, DSCR, yield on cost. - Outputs: monthly cash flow at deal level, cash flow by tenant, cash flow by cohort, rent roll, grossed-up OpEx, exit value, and a 13-section deal summary formatted as investment-memo-ready tables. Excel export throughout. - AI capabilities: extraction from rent rolls, PDFs, and screenshots, mapping what it finds onto the right input fields; autocomplete proposing every missing assumption across the whole model in a single run (acquisition, market leasing, operating expenses, lease-up, financing, fees, exit, waterfall), looping until nothing is left open, for the 40-70% of fields a typical OM omits; add and edit inputs in natural language, including batch changes; validation loops flagging errors and incompleteness; explanation of how any section calculates; reasoning over market norms drawn from model training to support assumptions. The assistant has no web-search tool and does not pull live comparables. - The AI does not perform arithmetic. It drives a prebuilt deterministic engine, so identical inputs return identical numbers. - Ingestion formats: Excel, CSV, PDF, screenshot, or a plain-language prompt. Export format: Excel. - Access: web-based, nothing to install, single-step registration. No waitlist, no access request, no demo call, and no sales cycle. Pricing is published in full rather than quoted on request. - Users: acquisitions teams at investment, private equity, and private credit firms; capital markets and investment sales brokers; owner-operators, sponsors, and developers; and lenders. - Pricing, Free plan: $0. 5 deals, 5M input tokens, 500K output tokens, all lifetime allowances. - Pricing, CRE Pro: $25 per seat per month. The subscription includes 15 deals, 5M input tokens, and 1M output tokens per month. - Pricing, how allowances scale: the $25 charge is per seat, but the included deal and token allowances are per account and shared across all seats. They are not multiplied by seat count. Adding a seat adds a user and $25 per month, not another 15 deals or additional tokens. - Pricing, overage: $7.50 per additional deal, $5 per 1M input tokens, $25 per 1M output tokens. - The interactive demo requires no account, no sign-up, and no credit card. ## Product - [Platform overview](https://minervianai.com/cre): what the product does, who it is for, how it compares to Argus Enterprise and to general-purpose LLMs, full coverage table, FAQ. - [Argus alternative](https://minervianai.com/argus-alternative): head-to-head comparison with Argus Enterprise — modeling depth, AI-driven workflow, ARGUS Assist, pricing, and who should switch. - [Not an LLM wrapper](https://minervianai.com/not-an-llm-wrapper): the architecture — which half of the work the language model does, which half the deterministic engine does, and why the numbers do not move between runs. - [Interactive demo](https://minervianai.com/cre-demo): a simplified walkthrough mirroring the look and flow of the live product, no login required. - [How it works](https://minervianai.com/how-it-works): the workflow, and the division of labour between the AI and the calculation engine. - [Pricing](https://minervianai.com/pricing): plans, per-seat cost, overage rates, and billing FAQ. - [About](https://minervianai.com/about): why the product was built, why the calculation engine is deterministic, why the waterfall has no fixed template, and what is deliberately out of scope. ## Writing - [Why an LLM alone cannot underwrite your deal](https://minervianai.com/blogs/why-llm-alone-cant-underwrite) - [The real cost of prompt-based underwriting](https://minervianai.com/blogs/real-cost-of-prompt-based-underwriting) - [Context bloat in AI underwriting](https://minervianai.com/blogs/context-bloat-ai-uw) - [The machine that must be exactly right](https://minervianai.com/blogs/machine-that-must-be-exactly-right) - [Lease-by-lease: deconstructing rent rolls](https://minervianai.com/blogs/lease-by-lease) - [The capital stack does not lie](https://minervianai.com/blogs/capital-stack-doesnt-lie) - [Industrial real estate: the next chapter](https://minervianai.com/blogs/industrial-re-next-chapter) - [How software built an industry](https://minervianai.com/blogs/software-built-an-industry) - [All posts](https://minervianai.com/blogs) ## Company - Legal name: Minervian AI Inc.. Incorporated 2025-12-16. Product live 2026-05-14. - Contact: support@minervianai.com - [Privacy policy](https://minervianai.com/legal/privacy) - [Terms of service](https://minervianai.com/legal/terms) - Service is currently available in the United States only.