AI for underwriting.
From the T12 to a price you can defend.
Drop the OM, the T12 and the rent roll. Every assumption is rebuilt, checked and questioned before you offer.
Three files in. An offer out.
An invented property, for illustration. Hover to hold a step; click a step to jump.
Documents
The T12 is mapped line by line and reconciled to its own totals. The rent roll becomes a unit mix. An offering memorandum's figures are listed for you to confirm, and never enter the model on their own.
Model
Unit mix, debt, growth rows, a ten-year pro forma, the waterfall. Blue cells are yours, everything else is formula, and every row shows how it is computed. It recalculates as you type.
Coach
Cells turn red, amber or green as you type. Each highlight carries the rule, the reason, and a one-click fix you can accept or overrule. What the model cannot prove becomes a question for the person who can answer it.
After the model, the work starts. Try it.
One at a time. Move a slider, press a lever: the live ones run the same engine as the workspace, on a sample 40-unit deal.
Set your own criteria. Every deal is measured against them, hard limits apart from preferences.
- Going-in cap rate6.56%≥ market 7.00%
- DSCR, lowest year1.16x≥ 1.25x
- Avg cash-on-cash1.7%≥ 7%
- IRR to investors-2.9%≥ 15%
- Equity multiple0.87x≥ 1.80x
- Cash flow every yearpositivenever negative
A second opinion on every assumption.
Never buy on someone else's numbers. The review ends in a verdict on each assumption, the price the corrected model supports, a stress test, and the questions to send before you offer.
Move the price. Watch the gates.
A 40-unit sample deal at a 7.0% market cap rate, 70% loan-to-value, 6.5% for 25 years. Nothing below is an estimate: every figure is recomputed by the model on each change.
Code computes. The coach explains. The two never trade places.
Bring the deal you are looking at this week.
Start from a blank model or drop the documents. The first review takes about as long as reading the OM.
