To analyze an offering memorandum (OM) quickly with AI, give the tool the OM together with the rent roll, the trailing-12 (T-12) operating statement and your model, and ask it to list every figure, recompute every formula and tie each OM claim to the document behind it, with the page or cell. In one finished run on a 120-unit multifamily package, the three files held 1,453 figures, formulas and statements. All of them were listed within 7 seconds of upload, and the finished review, with 633 claims marked and placed, arrived 49 minutes after upload. It found that the OM's NOI, occupancy, average rent and expense ratio did not match the documents the OM named as its source.
Written by the Titleman team with AI assistance (Claude). Reviewed by the Titleman review desk. The worked example is one finished Deal Review run on a Titleman QA account, uploaded at 07:49 UTC on October 15, 2026 and read on October 16, 2026. It reviewed a sample package we prepared ourselves (an offering memorandum, a T-12 with rent roll, and a buyer's underwriting model) with no client data and no address. Every figure below is from that run or is arithmetic on its figures. This is a method for analyzing an OM, not an appraisal, investment advice or a recommendation to buy or not buy.
The question
An OM is the seller's case for the deal, so the first read is fast and the risk is in the numbers. The question a buyer or lender wants answered quickly is not "what does the OM say" but "which of the OM's numbers hold up against the rent roll, the T-12 and my own model, and where exactly do they not?" A summary answers the first question. Only a check answers the second.
What the run was given and what it returned
| File | Offering memorandum | T-12 and rent roll | Underwriting model |
|---|---|---|---|
| What it is | Seller's OM, 3 pages | Rent roll of 120 units (September 30, 2026) and T-12 for the twelve months to the same date, 2 sheets | Buyer's base case, 1 sheet |
| Figures, formulas and statements found | 39 | 1,330 | 84 |
Across the three files the run found 1,453 items and checked all 1,453. Code recomputed 568 formulas and the cross-file ties, and the model read the remaining 879 items. Of the 429 items rated high materiality, 429 were checked. The review marked 633 claims: 592 verified, 34 to review and 7 red flags, and it placed all 633 on their page or cell.
The timing, from the run's own log:
- 7 seconds: the files were uploaded at 07:49:06 UTC and the full list of 1,453 items was ready at 07:49:13 UTC.
- 49 minutes: the finished review was delivered at 08:38:19 UTC.
OM claims against the documents behind them
The OM said on page 3 that "all figures are from the seller's rent roll dated September 30, 2026 and the trailing-12 statement for the same period." The run marked that claim red, because four headline figures differed from those two documents.
| What the OM or model said | Where | What the documents behind it showed |
|---|---|---|
| NOI USD 1,312,400 | OM | USD 1,029,089 on the T-12 |
| 96.0% occupied | OM | 91.7% on the rent roll (110 of 120 units) |
| Average in-place rent USD 1,515, above the USD 1,495 market average | OM | USD 1,458.23 on the rent roll, below market |
| Expense base "under 42%" of effective gross income | OM page 3 | 48.3% on the T-12 |
| Debt service coverage 1.27x | OM page 2 | 0.99x on the T-12 NOI |
| Loan USD 15,120,000 | Model cell B13 | USD 14,040,000 from the cell's own formula (65% of price) |
| Rent growth 3% a year | Model cell B21 | Market forecasts of 0.5% (2026), 1.0% (2027) and 2.3% (2028) |
The arithmetic behind each row, and how to run the same ties by hand, is on how to check the numbers in an offering memorandum.
Five checks for any AI tool that analyzes an OM
- Does it read the documents behind the OM? An OM's NOI and occupancy are claims about the T-12 and the rent roll. In the run, the NOI gap of USD 283,311 and the occupancy gap of 4.3 points were only visible because the T-12 and the rent roll were read with the OM.
- Does it recompute formulas, or read saved values? The model's loan cell was saved at USD 15,120,000 while its own formula gave USD 14,040,000. A tool that reads saved values carries the wrong loan into every return.
- Does it place every claim on a page or cell? A flag you cannot find in the file costs you the time the tool was meant to save. The run placed 633 of 633 claims.
- Does it say how much it checked? Ask for two numbers: items found and items checked. The run reported 1,453 and 1,453, with 0 unchecked.
- Does it list what it could not check, and can you read each flag? Count the red flags, then read them. One of the seven compared October's T-12 tax line of USD 21,650 with the OM's annual USD 259,800, which is twelve times that figure, so that flag clears on reading. The other six held.
What the run could not establish
The run listed six things it could not check. No address was given in any file, so the records on the property (title, liens, court cases, permits, zoning, flood and taxes), the year built and site area, recorded sales and leases for comparison, competitors and new supply within three miles could not be checked. No lease files, rent ledger, general ledger or bank statements were provided, so individual rents, lease dates and the T-12's monthly entries were taken as given. The OM's 14.8% IRR and USD 1,402,000 pro forma came with no cash flows behind them. A quick analysis is only quick if it is honest about these limits, so write them down for your own deal too.
What this does not tell you
- That the seller acted in bad faith. A gap shows that two documents disagree. The explanation may be innocent.
- The right price. The checks tell you which numbers hold up, not what the property is worth.
- That every run takes 49 minutes. The timing is for this package of three files and 1,453 items. A larger package has more to check.
- A decision. The buyer, the lender and the committee decide.
For what an OM contains, see what an offering memorandum is and the offering memorandum guide. For the documents behind it, see what a T-12 is and how to read a rent roll.
To have every material figure in your own OM, rent roll, T-12 and model marked verified, to review or red flag, each on its page or cell, send one deal’s documents to Deal Review.
Frequently asked questions
How do you analyze a commercial real estate offering memorandum quickly with AI? Give the AI the OM together with the rent roll, the T-12 and your model, not the OM alone. Ask it to list every figure, recompute every formula, tie each OM claim to the document behind it and point to the page or cell. In one finished run on a 120-unit multifamily package, the three files held 1,453 figures, formulas and statements, all listed within 7 seconds of upload, and the finished review marked 633 claims 49 minutes after upload.
How long does an AI analysis of an offering memorandum take? It depends on how much is checked. In the worked run, listing every figure in an OM, a T-12 with rent roll and an underwriting model took 7 seconds. The finished review, with all 633 claims marked verified, to review or red flag and each placed on its page or cell, arrived 49 minutes after the files were uploaded.
Can AI analyze an offering memorandum without the rent roll and T-12? It can summarize the memorandum, but it cannot check it. The OM's headline numbers are claims about the rent roll and the T-12. In the worked run the OM stated NOI of USD 1,312,400 and 96.0% occupancy, while its own T-12 totaled USD 1,029,089 and its rent roll showed 110 of 120 units occupied (91.7%). Neither gap is visible from the OM alone.
What should you look for in an AI tool for offering memorandum analysis? Five things: it reads the rent roll, T-12 and model as well as the OM; it recomputes formulas instead of reading saved values; it places every claim on a page or cell; it reports how many figures it found and how many it checked; and it lists what it could not check. The worked run found 1,453 items, checked 1,453, placed 633 of 633 claims and listed six things it could not check.
Written by the Titleman team with AI assistance (Claude). Reviewed by the Titleman review desk. Figures come from one finished Deal Review run of a sample 120-unit multifamily package on a Titleman QA account (uploaded 07:49:06 UTC, items listed 07:49:13 UTC, review delivered 08:38:19 UTC on October 15, 2026). The NOI gap of USD 283,311, the occupancy gap of 4.3 points and the 7-second and 49-minute timings are arithmetic on the run's figures. A review is not an appraisal, investment advice or a recommendation.