What is using Claude to screen a multifamily offering memorandum? It is the practice of feeding a broker's offering memorandum (OM) to Claude and asking it to recompute the deal on actual numbers, flag the marketing assumptions that inflate value, and return a clear first-pass screen so you know in minutes whether a deal is worth a full underwrite. The offering memorandum is a sales document, and a good screen treats it as a set of claims to verify rather than facts to accept. Done well, this turns a flood of marketed deals into a short list worth your analysts' time, and it sits squarely inside our complete approach to AI multifamily underwriting.
Key Takeaways
- An offering memorandum is a marketing document, so screening with Claude means verifying the broker's stated cap rate against the actual T12 and rent roll, not trusting the headline number.
- The output of a screen is a go or no-go decision in minutes, with an underwritten value range you can compare against the asking price before committing analyst hours.
- Claude reliably flags the five recurring inflation points: pro forma market rents, loss-to-lease recapture, understated expenses, a removed management fee, and optimistic other income.
- Screening is upstream of underwriting and downstream of sourcing; it is the filter that decides which marketed deals earn a full model.
- The screen never replaces full diligence, but it kills bad deals fast and lets a small team review three times the deal flow with the same headcount.
Why the OM Screen Is Its Own Skill
Sourcing automation finds deals; full underwriting models the ones you pursue; and the OM screen is the distinct middle step that decides which marketed deals cross from one to the other. If you have already built a system to surface deals, for example a Claude Project for multifamily deal sourcing automation, the screen is what you run on each offering memorandum that system flags. It is deliberately shallower than underwriting and far faster, because its only job is to answer one question: does the real deal, on real numbers, look anything like the deal the broker is selling?
That question matters because the offering memorandum is written to present the asset at its best. The headline cap rate is often a pro forma cap rate built on rents the property does not yet earn. The expense load may exclude costs a new owner will absolutely incur. The screen exists to strip the marketing back to the in-place reality, quickly and consistently, so a buyer does not spend a day modeling a deal that never penciled.
What to Load and What to Ask Claude First
Give Claude the offering memorandum PDF and, when the broker provides them, the actual T12 and current rent roll. Claude's large context window lets you paste the full OM plus the financials in one place, which is what makes the recomputation reliable. Start with a single grounding prompt: extract the deal summary. Ask for units, year built, the asking price, the broker's stated in-place and pro forma NOI, the stated cap rate, and the price per unit. This gives you the broker's version of the deal in a clean table before you challenge any of it.
Then ask the question that drives the screen: recompute the cap rate using only the in-place T12 income and a normalized expense load, and compare it to the broker's stated cap rate. Net operating income (NOI) is gross revenue minus operating expenses, and it excludes debt service and capital expenditures, so make sure Claude is not quietly importing pro forma rents into the in-place figure. The gap between the broker's cap rate and the recomputed in-place cap rate is usually the entire story of the deal.
The Five Inflation Points Claude Should Always Flag
Across thousands of multifamily offering memorandums, value is inflated in a small, predictable set of ways. Instruct Claude to check each one explicitly. First, pro forma market rents: the OM projects rents above in-place, and the screen should ask how much of the value depends on rents the property has never achieved. Second, loss-to-lease recapture, where the broker assumes every unit resets to market on turnover at an aggressive pace. Third, understated expenses, the most common trick, where the expense ratio sits well below a realistic range for the market and asset; a 28 percent expense ratio on a 1980s garden complex deserves scrutiny when comparable assets run closer to 40 to 45 percent. Fourth, a removed or below-market management fee, since a passive buyer will pay full third-party management even if the seller self-managed for free. Fifth, optimistic other income from fees, parking, or utility reimbursements that the T12 does not support.
Each flag should carry a dollar impact and a corrected number, so the screen produces an honest in-place NOI and a defensible value range rather than a vague sense of unease. This is the same line-by-line discipline our team applies when we help acquisitions groups build first-pass screens, and it is where The AI Consulting Network most often finds the difference between an asking price and a fair price.
From Screen to Go or No-Go
The screen ends with a decision, not a model. Have Claude state the recomputed in-place NOI, apply a market cap rate for the submarket and vintage, and produce an as-is value range. Set that range against the asking price. If the gap is small or favorable and the value-add thesis is real, the deal earns a full underwrite. If the asking price only works on pro forma rents the buyer does not believe, the screen returns a no-go with the reasons attached, and the analyst moves on without losing a day. The deepest part of any screen is understanding what could kill the deal later, which is why pairing this with an AI multifamily deal contingency analysis of what can kill a deal makes the go or no-go far more honest.
The leverage here is real. A two-person acquisitions team that can screen a marketed deal in 15 minutes instead of half a day can review three times the volume and still send only the genuine candidates to underwriting. CRE investors who want to operationalize this at scale can reach out to Avi Hacker, J.D. at The AI Consulting Network to build a repeatable screening workflow around their buy box.
The 15-Minute Screen in Practice
In practice the screen runs as a tight loop. Drop the offering memorandum and any actuals into Claude, ask for the deal summary table, then ask for the recomputed in-place cap rate and the five inflation flags with dollar impacts. Add two quick sanity checks: the price per unit against recent trades in the submarket, and whether the broker's stated cap rate is labeled in-place or pro forma, because brokers frequently lead with the pro forma number without saying so. Within 15 minutes you have a corrected NOI, a defensible value range, and a one-paragraph rationale you can paste into your pipeline tracker. The point is repeatability: the same prompts, run the same way on every deal, produce screens an acquisitions lead can trust and compare across the week's flow rather than ad hoc reads that vary by whoever happened to look.
Where the Screen Hands Off to Real Underwriting
A screen is a triage, and its discipline is to stay shallow. It does not model a ten-year hold, build a financing waterfall, or run a sensitivity table; those belong to full underwriting once a deal clears the screen. The screen also depends on the quality of what the broker provides. If the rent roll is a snapshot image or the T12 is incomplete, Claude can flag the gap but cannot manufacture the missing data, so the screen output should always note what it could not verify. When the deal advances, the natural next step is to automate rent roll analysis with Claude Projects, which takes the same source documents far deeper. Authoritative rent and expense benchmarks from sources like the National Multifamily Housing Council help calibrate what a realistic expense ratio and rent level should be for a given market.
Frequently Asked Questions
Q: How accurate is Claude's recomputed cap rate from an offering memorandum?
A: It is only as accurate as the source data. When the broker supplies a real T12 and rent roll, Claude's recomputed in-place cap rate is reliable and easy to verify. When the OM gives only pro forma figures, the screen should clearly label the result as an estimate pending actuals.
Q: Can Claude read a scanned offering memorandum or rent roll image?
A: Claude can read many scanned PDFs and images, but accuracy drops on low-quality scans and complex tables. For a screen, request the native file when possible, and always spot check the units, rents, and totals Claude extracts before trusting the recomputation.
Q: How is screening different from underwriting?
A: Screening is a fast first-pass that answers whether a marketed deal is worth modeling, usually in under 20 minutes. Underwriting is the full multi-scenario model you build only after a deal clears the screen. Confusing the two wastes time on deals that never qualified.
Q: What is the single biggest mistake in OM screening?
A: Accepting the broker's stated cap rate. It is frequently a pro forma number built on rents the property does not earn and expenses a new owner cannot avoid. Recomputing on the in-place T12 is the whole point of the screen.