Underwrite: deal analysis
Drop in the property documents, drag the dials, and read the returns before lunch.
About 14 minutes · Raising capital
Underwrite analyzes a property deal (built for multifamily) the way a spreadsheet never could: you drop in the seller's documents, the model fills itself with every number traced to its source, and then you drag dials and watch the returns recompute live. One engine drives everything you see and everything you export, so no two views of the deal ever disagree.
1Start a deal from the documents
- Open Underwrite and press New deal. Drop in the rent roll, the T12 (the trailing 12 months of income and expenses), and the offering memo. Any common layout from any property software works.
- Let it read. The rent roll sets the true unit count and rents. The T12 brings the operating picture, including every side income line like utility reimbursements and parking. The memo fills the property name, address, and asking price.
- Land on a working model. Every figure shows where it came from; click a number and see the exact page or spreadsheet cell it was read from. Anything odd is already flagged.
The AI is instructed to return nothing it can't actually read. A "still needed" banner names exactly which inputs are missing (address, taxes, in-place rent, price) and shrinks as you fill them. You always know what's real and what's still yours to confirm.
2Drive the model
- Drag a dial. Price, rent growth, exit cap rate, vacancy, interest rate, renovation pace. Every number, the 10-year cash flow, and the charts recompute before you let go.
- Ask instead of iterating. Goal-seek answers questions in one tap: the highest price that still clears your target return, the price where the first-year debt coverage meets the lender's floor, the rent growth needed for a 2x multiple.
- Switch views by question. Six named views instead of twenty tabs: Sandbox, Capital Stack, Operations, Value-Add, Returns, Confidence. The key returns stay pinned in a rail that never leaves the screen.
Underwrite › Harmony Grove Apartments
LP IRR 18.2% · Multiple 2.1x · DSCR 1.31x
Recomputes as you drag
Live
Firewall: expense ratio 31% is below market
Repairs may be hiding in CapEx. Click for the fix.
Check
- The firewall
- Seven built-in checks for the tricks that inflate a deal: occupancy gaps between physical and economic, repairs booked as capital improvements, expense ratios below 35 percent, lumpy utilities, flat insurance in a hard market, lease-rollover cliffs, and rent outliers. Each flag explains the reason and the fix, and a rent-roll-versus-T12 disagreement must be resolved before the model runs.
- Scenarios and odds
- Base, Upside, and Downside live as tabs. Sensitivity heatmaps show which two assumptions matter most together, a tornado chart ranks what moves returns, and a Monte Carlo runs 5,000 paths for a best-to-worst outcome cone.
- Location intelligence
- Add the address and public data fills a panel: demographics, jobs, flood risk, fair-market rents. One button runs a deep location analysis with population, income, rents, crime, and real cited news about the property itself, saved to the deal.
- Debt
- Model a fixed note or a floating-rate loan with an index, spread, and optional cap. Type in each lender's quote and the whole deal re-solves per quote, so the debt decision argues itself.
- The waterfall
- Split returns between investors and the sponsor with a real waterfall: preferred returns, catch-ups, multi-hurdle promotes, even A/B/C share classes. Every dial moves both sides' returns live, and the math is proven against an independent implementation to the cent.
- The validations panel
- The model checks its own math in front of you: sources equal uses, every month sums to the year, every dollar of distributions lands somewhere. Green rows, visible always.
3Share it, freeze it, hand it off
- Export. A branded PDF memo for humans, a values workbook for spreadsheets, or a full Excel model whose formulas actually recalculate offline and reproduce the engine to the cent.
- Freeze for the committee. Freezing stamps an immutable snapshot and locks edits, so nobody nudges approved numbers. The packet shows exactly what changed since the last review. Duplicate the deal to keep iterating.
- Turn the model into the raise. A frozen underwrite creates a draft TenthYards raise with the returns, structure, and property already filled in. After the deal closes, actuals post back against the underwriting, line by line.
Underwrite › Harmony Grove › Confidence
Frozen for investment committee
Snapshot v3 · Aug 21 · every edit locked
Frozen
PDF memoExcel modelCreate TenthYards raise
Every offering memo you parse and every deal you close quietly deposits its comps into your own private library. A year in, your own market memory beats the data vendors.
