AI Underwriting for CRE: Rent the Model, Own the Workflow

Blackstone and Brookfield just told the market they’d rather build their own AI than rent it off the shelf. Most of the coverage filed that under bad news for proptech. I read it the other way, and it changes where you should actually be spending.

- Sasha Deneux

The Take: the giants are building their own AI. Good. The model was never the edge.

Here’s what’s moving under the headlines. Blackstone and Brookfield are pouring money into custom AI. Anthropic and OpenAI have stood up enterprise arms pointed straight at big corporate clients, with CRE underwriting and portfolio work named out loud. Brookfield alone put 500 million dollars into OpenAI’s deployment company last month. So yes, the platforms that ran the last decade of proptech have a reason to sweat.

But look at what the giants are actually buying. They’re not trying to out-train Anthropic. Nobody is, and nobody should. They’re treating the model itself as a rented commodity and putting the real money into the part that compounds: how that model plugs into their deal process, their data, the way they price an asset.

That’s the lesson for everyone who isn’t Blackstone. You won’t own a better model than the shop down the street, and you don’t have to. What you can own is the process. The screen that matches your buy box. The underwrite that reflects how you actually price risk. The memo your committee reads without flinching. Rent the brain. Own the workflow.

Most of the operators I talk to already feel this in their gut. The ones pulling ahead with AI this year aren’t the ones with the slickest tool. They picked one ugly, repetitive job and made AI run it their way, start to finish. So let’s take one apart.

The Teardown: a first-pass underwrite, the way it should actually work

Picture the most thankless hour in acquisitions. A fresh OM, a rent roll, a T-12, and twenty to thirty minutes of pure data entry before anyone has an opinion worth having. At ten or fifteen deals a week, that’s most of a working day lost to typing instead of thinking. Here’s how you hand off the typing and keep the thinking.

First, ingestion. Feed it the rent roll and the T-12 and tell it to pull the line items into a clean structure. Nothing else. No interpreting yet.

Then the honest NOI. Have it rebuild in-place income the way you would if you had the hour: drop the model and employee units, strip the concessions, reset the tax line to your basis instead of the seller’s, normalize the expenses. Now the headline number and the real one sit side by side.

Then a first-pass return, off your assumptions, not its. Cap rate, cash-on-cash, and a flag everywhere the two NOIs disagree.

And here’s the part people skip, the one that turns a handy tool into a real liability. A model will fill an empty cell with a confident, plausible number every single time you let it. In an underwrite, a made-up rent is worse than a blank one, because it looks finished. So you hand it one hard rule: use only what’s in these documents, and where a number is missing, write a CONFIRM tag that names exactly what to go check. Never a guess.

What comes back is a first-pass underwrite in minutes, every figure either sourced or flagged for diligence, and the call, pursue or pass, still entirely yours. The machine typed and added. You decided. That’s the trade you want.

Signal

CRE AI pilots went from under 5% to 92% of firms in three years (JLL). Why it matters: the AI question at your committee isn’t if anymore. It’s which workflow goes first.

Wall Street is building, not buying. Blackstone and Brookfield are funding custom AI while Anthropic and OpenAI launch enterprise ventures (Bisnow). Why it matters: build-or-buy just became a live boardroom question well below the mega-cap tier, and it tells you the moat is your process, not the model.

EliseAI is live with 600-plus multifamily operators, including 75% of the NMHC Top 50. Why it matters: in multifamily, AI operations stopped being a pilot. They’re the floor now.

Bonside announced a 100 million dollar strategic partnership with i80 Group for an AI scorecard that underwrites brick-and-mortar tenant credit, with Kimco and Nuveen as early backers. Why it matters: AI underwriting is creeping from the asset to the tenant, which quietly changes how you read retail and mixed-use risk.

Primer now maps offering memos, rent rolls, and T-12s straight into your underwriting model. Why it matters: the ingestion grunt work from the teardown above is going commodity, so your edge moves up to judgment.

From NextAutomation

That underwrite teardown is the kind of engine we build and run for acquisitions teams. Your screening and underwriting logic, wired into your stack, with the no-guessing rule baked in so every number is either sourced or flagged. If first-pass underwriting is eating your analysts’ week, it’s the first thing worth automating. Book a call and we’ll find where AI pays back fastest in your pipeline.

That’s the read this week. Rent the model, own the process, and make the AI tell you what it doesn’t know.

Next week: the off-market sourcing engine, and how to score which owners are most likely to sell before they ever bring a deal to market.

- NextAutomation Team

Turn this play into a system

You can run this underwrite by hand, or let NextAutomation build the version that runs itself. Start with the free rent roll and T-12 normalizer, walk the live underwriting demo to watch it flag every number it cannot verify, and see how AI deal sourcing turns a one-off prompt into a system your pipeline runs on.

Keep reading

More on AI in the deal process: walk through an AI underwrite of a multifamily deal, the best AI tools for CRE underwriting, and (next week’s teaser) score which owners are most likely to sell off-market. From the archive: turn one walkthrough voice memo into a full content stack. Now live: the full off-market owner-propensity playbook.