Underwriting in commercial real estate is the disciplined process of answering one question: at this price, does this deal do what I need it to? It is the work of estimating a property’s income, its costs, its risks, and its likely return, then deciding whether the number the seller is asking makes sense. The word comes from insurance and lending — an underwriter is the person who decides whether a risk is worth taking, and on what terms — and it carried into real estate because buying a building is the same kind of bet. You are pricing a stream of future income against everything that could go wrong with it. This piece explains what that actually means for a small firm, the handful of numbers that matter, and where AI genuinely helps versus where it will confidently lead you into a bad deal.
What underwriting means in plain terms
Strip away the jargon and underwriting is estimating what a property will earn, what it will cost to own and run, and what those two facts are worth to you today. A building is a promise of future rent. Underwriting is the work of judging how reliable that promise is and pricing it accordingly.
In practice that means building a picture of the property’s income and expenses, usually a year or several years forward, and testing it against the price. If a retail strip center is listed at $3 million and, after honest accounting for rent, vacancy, taxes, insurance, maintenance, and management, it throws off $180,000 of net income a year, underwriting is what tells you whether that $3 million price is a fair one, a steal, or a trap. The number is not handed to you. You build it from the property’s records, the market around it, and a set of assumptions about the future — and the quality of those assumptions is the quality of your answer.
The reason the insurance word stuck is that both jobs are the same in spirit: someone has to look hard at a risk and decide what it is worth. An insurer underwrites a policy. A lender underwrites a loan. A buyer underwrites a purchase. In each case, “underwriting” means the careful, numbers-first judgment that happens before money is committed.
Lender underwriting vs. buyer underwriting
The single most common source of confusion is that “underwriting” describes two related but different jobs, and most explanations only cover one.
Lender underwriting asks: will this loan get paid back? A bank or agency lender looks at the property’s income, the size of the loan, and the borrower, and decides whether the deal is safe enough to lend against and on what terms. Their focus is downside protection — they want to be confident the building can cover its debt payments even if things soften.
Buyer or equity underwriting asks: will this deal earn the return I need? A purchaser or investor looks at the same property but from the other side of the table. Their focus is the upside and the risk to it — whether the price supports the return they are trying to hit, given what they will have to put in and what they expect to get out.
The two overlap heavily because they read the same documents and share several metrics, but they are not the same exercise. A deal can pass a lender’s test and still be a poor buy, and a great buy can be hard to finance. If you are a broker, an acquisitions shop, or a principal buying for your own account, the underwriting that matters most is the buyer’s version — pricing the deal against your own return, not just confirming a loan will clear.
Where underwriting sits in the deal
Underwriting is a middle step, and knowing what comes before and after it keeps the work from sprawling.
Before underwriting comes screening — the fast pass that separates the handful of deals worth real analysis from the flood that is not. Screening is a triage: does this property fit the mandate, is the asking price in a sane range, is it worth an hour of anyone’s time? Most deals die here, and they should. The mechanics of that first filter are their own discipline, covered in how investment firms filter opportunities before they ever build a model.
Underwriting is what a deal earns by surviving screening. It is the deeper look: build the income picture, run the metrics, test the assumptions, arrive at a value and a view on the price.
After underwriting, if the numbers work and you move toward an offer, comes due diligence — verifying that everything you assumed is actually true. Reading the real leases, inspecting the roof, confirming the taxes, checking the title. Due diligence is where you spend real money and time, which is exactly why you underwrite first: you do not want to be paying for inspections on a deal the math already killed. Screening filters, underwriting decides, diligence confirms.
The numbers that actually matter
A handful of metrics carry most of the weight in commercial underwriting. None of them is complicated on its own; the skill is in feeding them honest inputs.
Net operating income (NOI) is the property’s income after operating expenses but before debt payments and income taxes. Rent and other income, minus vacancy, minus the costs of running the building (taxes, insurance, maintenance, management, utilities you cover). NOI is the foundation — almost every other number is built on it.
Cap rate (capitalization rate) is NOI divided by price, expressed as a percentage. A property earning $180,000 of NOI priced at $3 million has a 6% cap rate. It is the quickest read on how a property is priced relative to its income, and it is how buyers compare one deal to another. For a plain-English walk through what a cap rate is and how to sanity-check the one a seller quotes, see the guide to cap rates and how AI helps pressure-test them.
Debt service coverage ratio (DSCR) is NOI divided by the annual loan payment. It tells a lender how much cushion the income has over the debt. A DSCR of 1.25 means the property earns 25% more than it owes the bank each year; agency and bank lenders commonly want to see a minimum in the 1.20–1.25 range, though it varies by property type and lender. Below their threshold, the loan shrinks or the deal does not finance.
Loan-to-value (LTV) is the loan amount divided by the property value — how much of the purchase is borrowed versus equity you bring. Cash-on-cash return is the annual pre-tax cash flow divided by the cash you actually invested, a plain read on what your money earns each year. And internal rate of return (IRR) rolls the whole hold — the cash flows plus the eventual sale — into a single annualized return figure, the number most equity investors ultimately judge a deal on.
You do not need all of them for every deal. A quick buyer’s read is often NOI, cap rate, and cash-on-cash; a financed deal adds DSCR and LTV; a longer hold adds IRR. The metrics are tools, not a checklist.
The assumptions are the whole game
Here is the point that most explanations bury: the math is the easy part. The hard part — the part that decides whether your underwriting is any good — is the assumptions you feed it.
Every number above rests on inputs you choose. What rent will the space actually command, and how fast will it grow? How much vacancy should you budget? What will taxes and insurance do after a sale reassesses the property? How much capital will the building need over your hold — a new roof, HVAC, tenant improvements to fill space? What cap rate will a buyer pay when you sell in five or seven years? Change any one of these and the return swings, sometimes dramatically.
This is why a clean spreadsheet can be dangerous. A model with a fabricated rent-growth number or a wishful exit cap rate produces a confident, precise, wrong answer — and precision reads as rigor even when the inputs are fiction. The discipline of good underwriting is being honest and conservative about the inputs, and stress-testing them: what happens to the return if rents grow slower, if vacancy runs higher, if you sell into a softer market? A deal that only works under optimistic assumptions is a deal that only works on paper.
The market view behind those assumptions is not guesswork you invent per deal — it is a thesis you should be able to defend. Where rents and vacancy in your submarket are actually heading is the same view a firm commits to when it publishes research, which is why building a defensible market read and underwriting deals against it are one discipline, laid out in the market-report framework for a lean firm.
Where the hours go for a lean firm
For a 4-to-20-person firm with no analyst to spare, the cost of underwriting is not the thinking — it is the grind around it. The judgment calls take a principal a few focused minutes. The hours disappear into everything before the judgment: pulling numbers off a rent roll, keying trailing operating statements into a model, chasing down comps, reformatting a broker’s PDF into something you can actually work with.
That imbalance is worth naming because it is where a small firm bleeds capacity. Every deal that reaches underwriting demands hours of mechanical data handling before anyone gets to the part that requires expertise. Multiply that across a pipeline and the firm underwrites fewer deals than it should, or underwrites them thinly, because the setup work exhausts the time. Principals end up making price decisions on half-built pictures, not because they lack judgment but because assembling the full picture costs more hours than the deal flow allows.
Understanding underwriting as two separable layers — the mechanical assembly and the human judgment — is what makes the next question answerable: which layer can you hand off, and which must stay with you?
Where AI helps — and where it must not
AI is genuinely useful in underwriting, but only on one side of that line. It is very good at the mechanical layer and dangerous on the judgment layer, and the whole value depends on keeping the two straight.
On the mechanical side, a general assistant or a purpose-built tool can read a rent roll and pull the tenants, rents, and lease dates into a table; summarize a trailing operating statement; reformat a messy PDF into structured figures; and do the arithmetic once you supply the inputs. This is the layer that eats a lean firm’s hours, and handing it off is where the time comes back. A small shop that automates the assembly can underwrite far more deals with the same headcount — the pattern that lets small firms out-operate larger ones is the argument of the small-firm playbook for out-operating institutional competitors.
On the judgment side, the line is firm: AI does not own your assumptions or your price. A language model asked for a rent-growth rate or an exit cap rate will produce a plausible-sounding number it has no basis for, and in an underwriting model a confidently wrong figure prices a bad offer. The rule is the same one that governs every safe use of these tools in deal work: the model reads and reconciles, but a human forms the assumptions, verifies every extracted figure against the source document, and decides the number. Used that way, AI turns underwriting from a slow, hours-heavy grind into a fast first pass a principal can then apply judgment to.
One more guardrail for a firm with no IT department: a rent roll and operating statement are confidential financial data. Before you put a deal’s documents into any AI tool, get in writing where the data is stored, whether your inputs train shared models, and how you delete your history. Major providers state that business-tier and API data is not used for training by default, but the contract is your safeguard, not the marketing copy.
For how underwriting fits alongside screening, comps, and the market view in a single deal-analysis workflow, the deal-analysis playbook for lean teams sequences the whole stack.
FAQ
What is underwriting in commercial real estate, in simple terms?
Underwriting is the process of estimating what a property will earn, what it will cost to own, and what those facts are worth today — then deciding whether the asking price makes sense. You build a picture of the income and expenses, run a few key metrics like net operating income and cap rate, test your assumptions, and arrive at a view on the deal. The word comes from insurance and lending, where an underwriter judges whether a risk is worth taking. Buying a building is the same kind of bet: you are pricing future rent against everything that could go wrong with it.
What is the difference between lender underwriting and buyer underwriting?
Lender underwriting asks whether a loan will get paid back; buyer underwriting asks whether the deal will earn the return the buyer needs. A lender focuses on downside protection — can the building cover its debt if things soften. A buyer focuses on the price relative to their target return. They read the same documents and share metrics, but a deal can pass a lender’s test and still be a poor purchase. If you are buying for your own account or advising a buyer, the buyer’s version — pricing the deal against your return — is the one that matters most.
What metrics are used in commercial real estate underwriting?
The core metrics are net operating income (NOI), cap rate, debt service coverage ratio (DSCR), loan-to-value (LTV), cash-on-cash return, and internal rate of return (IRR). NOI is income after operating expenses but before debt. Cap rate is NOI divided by price. DSCR is NOI divided by the annual loan payment. LTV is the loan divided by the value. Cash-on-cash is annual cash flow divided by cash invested, and IRR rolls the whole hold into one annualized return. You rarely need all of them on every deal — the set you use depends on whether the deal is financed and how long you plan to hold.
How is underwriting different from due diligence?
Underwriting decides whether a deal is worth pursuing; due diligence confirms that everything you assumed is actually true. Underwriting comes first — you build the income picture, run the metrics, and form a view on the price. If the numbers work and you move toward an offer, due diligence follows: reading the actual leases, inspecting the building, confirming taxes and title. Due diligence is where you spend real money and time, which is exactly why you underwrite first. You do not want to be paying for inspections on a deal the math already ruled out.
How long does it take to underwrite a commercial deal?
It varies from under an hour for a quick buyer’s read to several days for a complex, multi-tenant asset with a full multi-year model. For most small firms, the judgment itself is fast — a principal can form a view in minutes once the picture is built. The time sinks into the setup: pulling figures off a rent roll, keying operating statements into a model, gathering comps, and reformatting documents. That mechanical work, not the thinking, is what limits how many deals a lean team can underwrite in a week.
Can AI do commercial real estate underwriting?
AI can do the mechanical parts of underwriting well and should not be trusted with the judgment. It can read a rent roll into a table, summarize an operating statement, reformat a messy PDF, and run the arithmetic once you supply the inputs — the exact work that eats a lean firm’s hours. What it cannot do is own your assumptions or set the price. A model asked for a rent-growth or exit cap-rate figure will invent a plausible one, and in an underwriting model a confident wrong number prices a bad offer. The safe pattern is AI for assembly, a human for every assumption and the final number.
What are the biggest mistakes in underwriting a deal?
The biggest mistake is optimistic assumptions dressed up as a precise model. A clean spreadsheet built on wishful rent growth, understated vacancy, or a too-tight exit cap rate produces a confident, wrong answer — and the precision hides the fiction. Other common errors are forgetting that taxes reassess after a sale, ignoring the capital a building will need over the hold, and skipping a stress test. Good underwriting is conservative about inputs and asks what happens to the return if rents grow slower, vacancy runs higher, or you sell into a softer market.
Is it safe to put deal documents into an AI tool?
Only under the right contract. A rent roll and operating statement are confidential financial data, so before uploading them to any AI tool, get three answers in writing: where the data is stored, whether your inputs train shared models, and how you delete and export your history. Major providers state that business-tier and API data is not used for training by default, but for a firm with no IT department the contract is the safeguard, not the marketing copy. Treat a vague answer as a reason to keep proprietary deal data out of that tool.
Do I need to be a modeling expert to underwrite deals?
No. You need to understand what the metrics mean and be honest about the assumptions behind them — the math itself is simple arithmetic. Full financial modeling is a skill worth having for complex deals, but a principal can underwrite most straightforward properties with NOI, cap rate, and cash-on-cash, provided the inputs are real. The judgment about rents, vacancy, capital needs, and exit pricing matters far more than model sophistication. A simple model on honest assumptions beats an elaborate one on wishful ones every time.
Key takeaways
- Underwriting in commercial real estate is the disciplined answer to one question: at this price, does this deal do what I need it to? You build a picture of income, cost, and risk, then price it against the ask.
- Lender underwriting asks whether a loan gets repaid; buyer underwriting asks whether the deal hits your return. If you buy or advise buyers, the buyer’s version is the one that matters.
- Underwriting sits in the middle of the deal: screening filters the flood, underwriting decides, due diligence confirms before you spend real money.
- The core metrics — NOI, cap rate, DSCR, LTV, cash-on-cash, and IRR — are simple arithmetic. The hard part is the assumptions you feed them; a clean model on bad inputs is a confident wrong answer.
- AI is genuinely useful for the mechanical layer — reading rent rolls, summarizing statements, running the math — and dangerous on the judgment layer. Let it assemble; keep every assumption and the final number with a human.
Want to know where your firm’s hours actually go in deal analysis, and which parts AI can safely take off your plate? A short assessment maps that against your deal flow, your documents, and where AI belongs without introducing risk. Book your free AI-readiness assessment →
Arthur Wandzel
