Skip to main content

Commercial real estate AI

CRE AI: what it is and how commercial real estate teams use it

The plain guide to CRE AI: what the term means, the deal workflows it actually runs, what separates it from a general chatbot, and how commercial real estate teams learn it. Written for the people who have to defend the numbers in front of committee.

Get your free AI assessmentCRE AI training

The short answer

CRE AI, defined

CRE AI is artificial intelligence applied to commercial real estate work: reading offering memorandums, abstracting rent rolls and T-12s, screening and underwriting deals, drafting investment committee memos, and reporting to LPs. What separates it from a general chatbot is the setup around the model: it works from your actual deal files, follows your firm's underwriting rules and templates, and is built to show its work, the rent it used, the page it read, the assumption it applied, so a person can verify each number. At PSV we build CRE AI on Claude by Anthropic, chosen for how it handles long, rule-bound deal documents across a full session.

The core workflows

What CRE AI actually does on a deal

The value is not speed alone. It is that each output is traceable to the document line it came from, which is what makes AI usable on a real deal where a wrong number costs money.

  1. 01

    Deal sourcing

    Surface and shortlist on-market and off-market opportunities against a defined buy box, so what reaches a human is already filtered.

  2. 02

    Underwriting

    Extract the rent roll and T-12, build the operating model in your own template, and flag anomalies for review before they reach committee.

  3. 03

    OM and BOV review

    Pull the terms, rents, and assumptions that matter out of an offering memorandum and check them against your comps.

  4. 04

    IC memos

    Turn a completed model into a committee-ready narrative: thesis, risks, comps, sensitivities, and a recommended bid range.

  5. 05

    LP reporting

    Assemble quarterly investor updates from source data, drafted in your voice and traceable back to the statement behind each figure.

CRE AI vs a generic chatbot

The difference is the setup, not the chat box

A generic chatbot

  • Answers from general knowledge.
  • Will guess at a number when the document is ambiguous.
  • Is not bound to your underwriting rules or template.

A CRE AI setup

  • Grounded in your actual deal files.
  • Constrained to your underwriting rules and templates.
  • Built to show its work: the rent it used, the page it read, the assumption it applied.

That difference, cited-truth as a system property rather than a lucky answer, is what makes AI safe to use where the numbers reach committee.

Who builds it, and how to learn it

Built by operators and AI engineers, for operators

PSV (Pacific Software Ventures) builds CRE AI two ways. The CRE AI Institute teaches operators to run their own work AI-native, and PSV's advisory and development team builds custom agents inside a firm's environment and stays on as its embedded AI team. The firm owns what PSV builds for it. PSV is led by Zed Truong, who previously worked in institutional real estate at PGIM Real Estate and in brokerage at Voit Real Estate. He has closed $1B+ in commercial real estate transactions and underwritten $11B+ across asset classes. He works alongside an ex-Goldman Sachs engineer and AWS AI Scholar, an infrastructure engineer out of Kochava and HP Enterprise, and an advisor who was Uber's seventh employee.

Why the Company Brain matters here

The whole deal, run the way your firm already runs it.

Two firms can license the same model, the same data, and the same tools by Friday. What one of them has and the other does not is its own accumulated judgment: the calls a partnership has already made, the standards it settled years ago, the deals it walked away from and the reason why. That is the line between a firm running AI and a firm running its own AI.

  • Your firm can answer questions about its own history.

    Have we looked at this sponsor before, what did we conclude, what did we price it at. In most firms that answer sits in three inboxes and one person's memory, so it costs a day to chase down and usually nobody bothers. A firm with its context settled treats the question as routine.

  • The advantage compounds instead of resetting.

    A general tool is the same on your four hundredth deal as it was on your first. A firm with institutional context behind the work is further along every quarter, because the year it just spent underwriting raises the floor for the year ahead.

  • Your edge stops being who bought the better tools.

    Any competitor can subscribe to the same data and the same software this quarter. Nobody can subscribe to two decades of calls your partners have made, and on a contested deal that is what makes one read worth more than another.

See how the Company Brain works

Learn it hands on

Learn to run your CRE work AI-native

The CRE AI Institute teaches the exact workflows above on a real practice deal, with a 7-day free trial and a 30-day money-back guarantee.

Start CRE AI trainingDeploy it across a firm

Frequently asked

CRE AI questions

What is CRE AI?
CRE AI is artificial intelligence applied to commercial real estate work such as deal sourcing, underwriting, offering memorandum review, investment committee memos, and LP reporting. What sets it apart from a general chatbot is that it works from your actual deal files, follows your underwriting rules, and is built to show the source behind the numbers it produces.
What does CRE stand for in CRE AI?
CRE stands for commercial real estate. CRE AI is the use of AI for commercial real estate workflows specifically, as opposed to residential or general-purpose AI.
How do commercial real estate brokers and investors use AI?
Brokers use AI to draft BOVs, pull comps, and abstract offering memorandums. Investors and acquisitions teams use it to screen deals against a buy box, abstract rent rolls and T-12s, build models in their own templates, and write IC memos. The common thread is grounding the output in real deal documents rather than general estimates.
Can AI underwrite a commercial real estate deal?
AI can handle much of the underwriting workflow: extracting the rent roll and T-12, populating an operating model in your template, and flagging anomalies. It works best as an operator-supervised assistant that shows its sources, so a person can verify the numbers before they reach committee.
Is Claude or ChatGPT better for CRE underwriting?
For long, rule-bound deal documents, PSV builds on Claude by Anthropic for how it handles long context and detailed instructions across a session. The larger point is that the setup around the model, grounding it in your files and rules, tends to matter more than the model brand.
How do I learn CRE AI?
The CRE AI Institute by PSV teaches operators to run their work AI-native through structured courses, live workshops, and a practice deal with answer keys. It starts with a 7-day free trial. Firms that want AI deployed across the whole team can work with PSV directly.
What AI tools do commercial real estate teams use?
Teams pair a general reasoning model such as Claude or ChatGPT with CRE-specific tools for data, comps, and pipeline, chosen by role: broker, investor, analyst, or developer. The differentiator is the grounded workflow around the tools, not any single tool on its own.

Keep reading

Learn it, see it working, or go deeper on one piece

PSV Research Desk

Related PSV analysis

Source-cited reporting on the workflows in this guide.

CRE AI · Weekly workflow teardown

One CRE workflow taken apart every week.

Sourcing, underwriting, OM review, IC memos, LP reporting. Each issue walks one of them end to end: what AI carried, what it got wrong, and what a person still had to check before the number left the building. Written for operators, not for a tools roundup.

No spam. Unsubscribe in one click.

Next step

Find out which of these workflows should run AI-native first.

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

Get your free AI assessmentTalk to PSV

We use cookies for analytics. Your data helps us improve the site.