Skip to main content

AI analyst for CRE · Updated November 21, 2026

The AI analyst for CRE, built on your own model

What an AI analyst does on a deal, what your human analysts still own, a day in the workflow, how to judge one, and your three ways to get one.

Disclosure

Pacific Software Ventures (PSV) builds custom AI analysts for CRE firms, so we have a stake in what you choose. The checklist below applies to any AI analyst, ours included, and the results link to the full case studies.

What is an AI analyst in commercial real estate?

An AI analyst in commercial real estate is an AI agent that does the analyst work behind every deal. It reads OMs, rent rolls and T-12s, screens each deal against your buy box, fills your own Excel model with every number tied to its source, pulls comps, and drafts the first IC memo and the quarterly LP report. It does not replace your analysts. They keep the assumptions, the market read, the checks and the recommendation, and spend their day reviewing instead of re-keying. The ones that work are built on your own model, criteria and templates. Silverado Interests' AI analyst works deal screening, underwriting prep and investor reporting from one email address the Dallas team forwards work to.

What does an AI analyst do in commercial real estate?

These are the analyst tasks firms we talk to ask for, in roughly the order a deal needs them. For each one, what the AI does and what your human analyst still owns.

  • OM to your own model. The AI reads the offering memorandum, rent roll and T-12 and fills your firm's own Excel model, formulas intact, with each input linked to the page or cell it came from. It does not swap your model for its own. Your analyst owns the exit cap, rent growth, vacancy, capital plan and every assumption the seller's pro forma tries to set for you. AI underwriting for CRE
  • Buy-box screening. Every broker email, teaser and OM that lands is checked against your written criteria: market, asset class, size, vintage, basis and return bar. Each deal comes back as pursue, pass or needs information, with the rule behind the call in one line. Your team owns the buy box itself and the decision on which deals get a second read. AI for CRE acquisitions
  • Rent roll and T-12 normalization. The AI turns the seller's rent roll and operating statement into a clean operating snapshot in your categories: unit mix, in-place vs market rent, lease expirations, and expenses mapped to your chart of accounts. It ties the rent roll to the income line and flags what does not tie. Your analyst owns the calls on one-time items, concessions and what counts as recurring. Free T12 analyzer
  • Comps for every deal. It builds the comp set from your own past underwriting, the comps in broker packages, and the public and licensed sources whose terms allow it, then lines the new deal up against them. Keep data vendors' exports out unless you have their written permission. Several vendors' terms bar putting them into outside AI tools. Your analyst owns which comps actually count for this asset.
  • The first IC memo draft. From the model and the deal documents, it drafts the memo in your house format: the deal, the market, the business plan, returns, and risks with mitigants. Every figure in the memo matches the model because it came from the model. Your deal team owns the thesis, the edits and the case they make to committee. AI investment committee memos
  • LP reporting and variance. After closing, the same analyst reads the monthly GL export from your property management system, builds budget vs actual for each asset, drafts the variance commentary and then the quarterly LP letter from those numbers. Your asset managers own the explanation they sign, and nothing reaches an investor before a person approves it. AI investor reporting
  • The pipeline tracker and deal memory. It keeps your deal tracker current from the email threads and answers questions like what you bid on that submarket last year, cited to your own files. Your team owns the broker relationships and the follow-up calls. How a Company Brain works

Does an AI analyst replace a junior analyst?

No. It replaces the part of the job nobody hired an analyst to do: re-keying OMs, rebuilding rent rolls, reformatting the T-12 and updating the tracker. What stays with people is the part that makes an analyst valuable.

  • The assumptions. Exit cap, rent growth, vacancy, capital and the debt terms you can actually get. An AI can show what the seller assumed and what your recent deals assumed. Choosing the number is a judgment about this asset in this market, and it stays with your team.
  • The market read. The site visit, the call to the leasing broker, the property manager who tells you why occupancy dipped last spring. None of that is in the OM, and it is often what decides the deal.
  • Checking the work. A good AI analyst shows every source so a person can check it in minutes. Someone still has to do that check, and a junior analyst who reviews many populated models learns faster than one who types in a few.
  • The recommendation. Whether to bid, at what price, and how to defend it in front of the committee. The AI drafts; your people decide and sign.

What does a day with an AI analyst look like?

Here is a typical day for an acquisitions team that works with one. The team reaches it the way they reach a person: by email, or in the tools they already use.

  1. Before the team logs on. Overnight broker emails and new OMs have been read and screened. The morning summary lists the deals that fit the buy box, the ones that missed and why, and the ones waiting on a rent roll or a price.
  2. Morning: a deal worth a look. The VP forwards the OM on a deal that fits, with a one-line note. Within the same thread the AI sends back your model filled from the OM, rent roll and T-12, with a list of the inputs it could not find and the places the seller's numbers do not tie.
  3. Late morning: the analyst reviews. The analyst opens the model, clicks through the source links on the numbers that matter, corrects a misread concession and sets the assumptions. This is the review that replaces the build.
  4. Afternoon: comps and the memo. The AI lines the deal up against your comp set and drafts the short IC memo for approval to bid. The analyst rewrites the thesis paragraph and adds what the leasing broker said on the phone.
  5. Month end: the hold side. When the books close, the same analyst builds budget vs actual for each owned asset and drafts the variance notes, and at quarter end the LP letter. Asset managers edit the variance notes, and a named person signs off the LP letter.
  6. End of day. The tracker is current, every deal has a status and a next step, and the team spent the day on decisions instead of data entry.

Examples: AI analysts at real CRE firms

Three of our clients and the analyst work their AI handles.

  • Silverado Interests, an AI analyst the team emails. Silverado Interests' AI analyst works deal screening, underwriting prep and investor reporting from one email address the Dallas team forwards work to. Read the case study
  • Westlake Realty Group, underwriting. Westlake Realty Group cut per-deal underwriting from 6-8 hours of senior time to a 20-minute review, in its own model with every number traced to source. Read the case study
  • Accomplish Collective, across the deal cycle. A real estate sponsor put four AI employees on deal sourcing, underwriting, asset management and investor reporting, inside its own Microsoft 365 tenant. Read the case study

Off-the-shelf AI analyst or one built on your own model?

The right choice depends on how much of your edge lives in your own model, criteria and templates. Here are the three ways firms get one, and who each fits.

  • Have one built on your model, buy box and templates (ours). We build the AI analyst around your own Excel model, written buy box, IC memo format and reporting templates, inside the tools your team already uses, then stay on as your AI team as the model, the criteria and the market change. It fits firms whose underwriting is their own and who want the output in their own files. Silverado Interests, for one, has an AI analyst with its own email address. Read the Silverado Interests case study
  • Do it yourself in Claude. Low cost to try, and it works for one analyst on one task if you write the criteria down and set it up to cite its sources. Your team maintains the prompts, the model links and the checks as they change. Claude for CRE operators
  • Buy an off-the-shelf AI analyst product. A product works the way its vendor designed it, in its own screens and usually its own model, at a per-seat price. It fits a team that is happy to adopt that model and workflow as its own.

How do you judge an AI analyst?

Judge it the way you would judge a new hire: on work you can check. These tests apply to any AI analyst, ours included.

  1. Run it on deals you already closed. Give it three or four deals from last year where you know the right answer. Compare its model, screen and memo with what your team produced. Where it disagrees, find out whether it misread a document or found something you missed.
  2. Every number should tie back to the source document. Click any input in the model and you should land on the page of the OM or the cell of the rent roll it came from. A number with no source is a number someone has to re-check by hand, which is the work you were trying to remove.
  3. It should cite sources and say what it could not find. A good analyst says the T-12 is missing March, or that the rent roll total does not match the income line. An AI that fills the gap with a plausible guess is worse than one that leaves it blank and tells you.
  4. Check that it ties out. Rent roll to income, unit count to the OM, model returns to the memo. These tie-outs are fast to check and catch most extraction errors before they reach a decision.
  5. Does it work in your model and your templates? If the output is in someone else's model, your team re-translates every deal back into yours. The test is whether your formulas still work after it fills the inputs.
  6. Where do the deal documents go? Ask where OMs and rent rolls are stored, who can see them and whether they are used for training. Keep data vendors' exports out unless you have their written permission. Several vendors' terms bar putting them into outside AI tools. Is AI safe for deal documents
  7. Who keeps it current when your model changes? Models gain tabs, buy boxes move with the market, memo formats change with a new committee. Someone has to update the analyst when they do, or it drifts out of step with how your firm actually works.

What does an AI analyst cost, and how long does it take to set up?

The useful comparison is not the price of software. It is the analyst and senior hours your team spends today on reading, re-keying and reformatting, and the deals you pass on because nobody had time to underwrite them.

  • Start from hours, not seats. Add up the time per deal your team spends getting an OM into the model, cleaning the rent roll and T-12 and drafting the memo, then multiply by your deal count. Westlake Realty Group cut per-deal underwriting from 6-8 hours of senior time to a 20-minute review, in its own model with every number traced to source. Read the Westlake Realty Group case study
  • What drives setup time. Three things: how complex your model is, how clearly your buy box is written down, and how consistent your memo and reporting templates are. A firm with a written buy box and one standard model is ready sooner than one with a model per analyst.
  • Start with one workflow. Start with the step that costs your team the most hours, often OM to model or deal screening, prove it on closed deals, then add the memo, comps and reporting to the same analyst.
  • Plan for upkeep. An AI analyst is not a one-time install. Budget for someone to keep it in step with your model and criteria, whether that is your own team or ours.

Common questions

What is an AI analyst in commercial real estate?
An AI agent that does the analyst work on a deal: reading OMs, rent rolls and T-12s, screening deals against your buy box, filling your model, pulling comps and drafting IC memos and LP reports, with every number tied to its source. Your team reviews the work and makes every decision.
Will an AI analyst replace junior analysts?
No. It takes the re-keying, reformatting and tracker upkeep off their plate. The assumptions, the market read, checking the work and the recommendation stay with people, and juniors learn faster reviewing many populated models than building a few by hand.
Can an AI analyst work in our own Excel model?
It should. A well-built AI analyst fills the inputs of your existing model with formulas intact and links each input to its source. If a tool needs you to move to its own model, your team ends up translating every deal back into yours.
How accurate is an AI analyst?
Accurate enough to trust only after you test it on deals you already closed and it ties out: rent roll to income, unit count to the OM, model to memo. Every number should link to its source, and it should flag what it could not find instead of guessing.
How does a team work with an AI analyst day to day?
The way they work with a person: forward a deal or a request with a short note and the work comes back. Silverado Interests' AI analyst works deal screening, underwriting prep and investor reporting from one email address the Dallas team forwards work to.
How much faster is underwriting with an AI analyst?
Westlake Realty Group cut per-deal underwriting from 6-8 hours of senior time to a 20-minute review, in its own model with every number traced to source.
Is it safe to give an AI analyst our OMs and rent rolls?
It can be, on the right plan with the right settings. Check where files are stored, whether they are used for training and who can see them. Keep data vendors' exports out unless you have their written permission. Several vendors' terms bar putting them into outside AI tools.
Should we buy an AI analyst product or have one built?
If your underwriting runs on your own model, buy box and memo format, build the analyst on those, because then the output lands in your own files and nobody retypes it into a vendor's format. Test either one on deals you already closed.

Next step

See where AI saves your firm time across your firm.

Ten minutes. An honest read on where you stand, the highest-value opportunities, and which engagement fits, if any.