How long does it take to underwrite a deal?
It depends almost entirely on how much of the work is re-keying rather than thinking. A first-pass screen on a clean multifamily package is an afternoon for somebody experienced. A full institutional underwriting with a lease-by-lease commercial rent roll, a debt quote and a partnership waterfall runs into days, and longer when documents arrive late. Most of that time is not analysis: it is reading an offering memorandum, a scanned rent roll and a trailing twelve month statement, typing them into a model, and then reconciling the three when they disagree.
Where the time actually goes
Split the work into the part a person has to judge and the part a person is only transcribing, and the split is uncomfortable. Unit mix, in-place rents, lease expiration dates and twelve months of expense lines are all facts sitting in a document that somebody retypes into a spreadsheet. Nothing about that step is analysis, and on a scanned or photographed page it is the slowest step in the process.
The judgment steps are the reconciliation between the rent roll and the trailing statement, the expense normalization, the market rent call, the debt sizing and the exit. Those are where an experienced underwriter earns the fee, and they are compressed or skipped when the transcription ran long.
| Stage | What it consumes | Transcription or judgment |
|---|---|---|
| Reading the package | Memorandum, rent roll, T-12, tax bill | Reading |
| Getting the numbers in | Unit mix, rents, expiries, expense lines | Transcription |
| Rent roll against the T-12 | Why annualized rent does not tie to collections | Judgment |
| Normalizing expenses | Taxes reassessed, insurance requoted, fee to market | Judgment |
| Sizing the debt | Three tests, at the lender's rate and schedule | Judgment |
| Exit and sensitivity | Exit cap, hold, and the two drivers that matter | Judgment |
| Producing the outputs | Workbook, memo, committee pages | Transcription |
A screen and a full underwriting are different jobs
A screen exists to decide whether a deal is worth a full underwriting. It builds a defensible year one, sizes indicative debt, and produces a bid range. Most deals stop here, which is the point: an acquisitions team passes on the large majority of what it looks at, so the screening cost lands on every deal in the pipeline while the fee only lands on the ones that close.
A full underwriting adds the things the screen approximated. Lease-by-lease rollover with renewal probabilities, tenant improvements and downtime. A real debt quote rather than an assumed constant. Partnership economics. After-tax treatment. Scenario and sensitivity work across the drivers that turn out to carry the outcome. Each of those is a genuine addition, and none of them is retyping.
What makes a deal slow
Document quality first. A native spreadsheet rent roll is minutes; the same rent roll flattened into a PDF is an hour; a photograph of a printed page is worse. A commercial rent roll with recovery structures, expense stops, options and escalations is slower again, because the lease terms behind each row have to be read rather than counted.
Then contradiction. A marketed vacancy that does not match the rent roll, a repairs line hiding a roof, a management fee below market, a tax bill that predates the last reassessment: each disagreement is a real finding and each one costs time to run down. That is time well spent and it is exactly the time that gets squeezed when the transcription overran.
Then waiting. A debt quote, an insurance quote and a tax estimate come from other people on their schedule, and a bid deadline does not move for any of them. Much of the elapsed time on a deal is not work at all.
What software changes, and what it does not
The honest claim is narrow. Software of this kind removes the transcription and the version drift. It does not remove the judgment, and a tool that claims to is claiming to make the market rent call for you.
Altyst reads the documents a deal already arrives with, in PDF, Excel, CSV, Word, plain text or a photograph of a page, including scanned pages through OCR, and it can start from a listing link and the asking price. Each extracted value is proposed with its source document and page and a confidence signal for your review, and a deterministic engine in exact decimal arithmetic computes the model from what you confirm. Changing any assumption recomputes returns, cash flow, debt and the downside case together, and clicking a figure shows the formula behind it. That is what the phrase on the front of this site, underwrite any property in seconds, is describing: the draft, not the decision.
The part that should still take time
Whether the market rent column is a market rent or an aspiration. Whether a renewal probability is evidenced by the tenant's history or chosen to make the deal work. What the exit cap rate should be, decided by policy before the model is built rather than solved for afterwards. Why the rent roll and the trailing statement disagree by $38,000. None of those is faster because the data entry was.
Speed on the mechanical half is only worth having if it buys more attention for this half. A model produced quickly and reviewed carelessly is a worse outcome than a slow one, and the reason to remove the retyping is that the retyping was never where the answer came from.
Related questions
How long does it take to underwrite a multifamily deal?
A first-pass screen on a clean package with a native rent roll and a trailing twelve month statement is an afternoon for an experienced underwriter. A full underwriting with normalized expenses, a real debt quote, partnership economics and sensitivity work runs into days. The variable that moves it most is not deal size but document quality, since a scanned or photographed rent roll can take longer to transcribe than the entire analysis takes to perform.
Why does underwriting take so long?
Because most of it is not analysis. Reading the offering memorandum, the rent roll and the operating statement and retyping them into a model is the largest single block of time, and it happens before any judgment can be applied. Reconciling documents that disagree is the second, and waiting on a debt quote, an insurance quote or a tax estimate is a third that is not work at all.
What is the difference between screening a deal and underwriting one?
A screen produces a defensible year one, indicative debt and a bid range, and exists to decide whether a deal deserves more work. A full underwriting adds lease-by-lease rollover, a real debt quote, partnership economics, after-tax treatment and scenario work. Most deals in a pipeline never move past the screen, which is why the screening cost lands on every deal a team looks at.
Can AI underwrite a deal faster?
It can remove the transcription, which is the largest block of time, and it is genuinely well suited to reading a scanned rent roll. It is poorly suited to performing the calculation, because a language model produces arithmetic by predicting text and can return a confident wrong figure. The reliable pattern is a model proposing values with their source, a person confirming them, and a deterministic engine computing the result.
How long should you spend on a deal you will probably pass on?
Enough to be confident in the pass, and no more. That is why the screen exists as a separate step with a lower standard of completeness than a full underwriting. The economics of an acquisitions pipeline are decided by how cheaply a team can reach a defensible no, because that cost is paid on every deal while the closing costs are paid on very few.
Keep reading