AI in CRE: What 17 Firms Want to Automate—and What Comes Next

AI in commercial real estate has moved past the curiosity stage, but most firms have not yet turned that interest into repeatable operating workflows.

We surveyed 17 senior executives at YPO-sized CRE firms to understand what their peers are doing with AI, where they want automation next, and what separates scattered experimentation from production. The clearest signal is not another chatbot. It is the race to build controlled workflows across underwriting, reporting, due diligence, and property operations.

Executive summary

• Peers are convinced, but most are not yet operational. 14 of 17 firms rate AI an 8–10 priority, while only 5 report production use beyond chatbots or copilots.
• Competitive attention is converging on workflows. Finance and reporting attract 15 firms, property and portfolio operations 14, and deal execution 13.
• The transaction process is the clearest near-term battleground. 13 firms want automation somewhere in acquisitions, underwriting, or due diligence; 9 selected at least two of those stages.
• Sponsorship separates activity from execution. Ten of 13 firms with strategic or actively pushing leadership have launched an initiative, versus 1 of 4 with a tentative posture.
• The executive mandate is practical. Pick a bounded workflow, assign an operating owner, preserve human judgment, and demand a 90-day before-and-after result.

Where peers stand

The peer group is in the messy middle: access and experimentation are common, but governed production workflows are not. The small production cohort is not limited to the largest firms; every respondent beyond copilots has a back office of 25 or fewer. That makes operating design—not scale—the more intriguing early signal.

Leadership posture is associated with execution

Ten of 13 firms whose leaders describe AI as strategic or actively push adoption have launched an initiative—a 77% launch rate. Only 1 of 4 firms with tentative or ROI-dependent leadership has done so. This small sample shows association, not causation, but it reinforces a practical point: visible sponsorship helps turn interest into an owned initiative.

Where peers are placing bets

The intrigue is in the gap between consensus and production. Most firms describe a connected automation agenda across finance, operations, and deals, yet only five have crossed beyond copilots. The near-term competitive question is not who has the most tools; it is who turns a recurring process into a controlled operating lane first.

The deal-workflow race

Acquisitions and underwriting are the clearest near-term battleground. Respondents are already using AI for research, underwriting support, memo drafting, lease and document review, and investment-committee preparation. What they want next is a continuous, governed process that can ingest deal materials, normalize facts, flag missing evidence, populate approved model inputs, draft decision materials, and route exceptions to a human reviewer.

What “automation” means in practice

1. Ingest: offering memoranda, rent rolls, leases, data-room files, comps, and approved underwriting assumptions.
2. Structure: classify documents, extract facts, reconcile entities, and surface missing or conflicting evidence.
3. Analyze: populate approved inputs, run bounded scenarios, and draft commentary without hiding assumptions.
4. Decide: assemble memo and IC materials, then route risks, exceptions, and judgment to accountable humans.

This is the natural next step after the use cases explored in Automating Deal Sourcing in Real Estate. The operating value comes from connecting adjacent steps—not merely making one prompt faster.

Three workflows worth operationalizing

These blueprints translate peer interest into an executive design brief. They are illustrative syntheses—not verbatim respondent plans—and deliberately keep human approval around assumptions, accounting treatment, sensitive communications, capital decisions, and external distribution.

Five firms, five strategies

Each vignette reconstructs one respondent’s reported maturity, workflow pattern, build approach, and next priorities. None is a composite or a direct quotation. Identifying details—including name, exact title, geography, and company identity—are suppressed.

Firm A: The workflow-first owner-operator

Current position. A lean owner-operator reports multiple live workflows and strong, expanding success using mostly off-the-shelf tools. Current activity spans research, underwriting, document review, reporting, meeting follow-up, and workflow handoffs.

What comes next. Connect acquisitions, due diligence, and legal/compliance into a more continuous deal process.

Peer lesson. A small team can reach production without beginning with a custom platform; the differentiator is connecting tasks into an owned workflow.

Firm B: The bottom-up experimenter

Current position. Employees already use AI across research, investment memos, underwriting, document review, reporting, meetings, CRM administration, and handoffs, yet the firm reports no formal initiative or firm-level ROI claim.

What comes next. Priorities include acquisitions, underwriting, due diligence, leasing, and reporting.

Peer lesson. Broad employee use can create a shadow operating system: useful activity without standard ownership, controls, or measurement.

Firm C: The custom operating integrator

Current position. A property operating platform reports multiple live, mostly custom workflows and clear ROI in limited areas, connecting reporting, pipeline administration, stakeholder communications, and handoffs.

What comes next. Extend the model into internal operations, compliance, investor relations, and reporting.

Peer lesson. Custom integration appears most useful where fragmented operating data and recurring communications must move through one controlled process.

Firm D: The connected investment workflow

Current position. A lean investment manager reports limited production use, mostly custom, with clear ROI in selected areas. Its current pattern connects investment memos, underwriting, document review, IC preparation, reporting, reconciliation, and handoffs.

What comes next. Carry the same structured approach into asset management, internal operations, diligence, and compliance.

Peer lesson. The next advantage may come from extending a transaction data model beyond closing rather than starting a separate post-close AI effort.

Firm E: The scaled platform still at copilots

Current position. A larger integrated platform reports broad employee use and clear ROI in limited areas, but its initiative remains at the chatbot/copilot stage rather than production workflows.

What comes next. Priorities include underwriting, finance/accounting, reporting, internal operations, and marketing/business development.

Peer lesson. Scale and broad tool access do not automatically create production; integration, operating ownership, and controls remain separate work.

Together, these firms reinforce the distinction described in The Real AI Divide Is Between Experimentation and Operations: access creates activity, but workflow ownership creates operating leverage.

A 90-day executive agenda

The goal is not a firmwide AI transformation. It is a controlled decision about whether one high-value workflow can move from scattered assistance into a repeatable operating lane.

Days 1–30: Define the lane

Choose one high-volume workflow, map its inputs and handoffs, establish the baseline, and name both a senior sponsor and a day-to-day operating owner. Baseline cycle time, rework, external spend, capacity, and decision turnaround.

Days 31–60: Prove the workflow

Run a bounded pilot using approved data and tools. Keep judgment, exceptions, and external output behind explicit human review gates. The proof is repeated use by the target team, not a one-time demonstration.

Days 61–90: Decide on production

Integrate only what earned it. Document controls, train users, compare results with the baseline, and stop or redesign workflows that cannot show repeatable value. The decision is simple: scale, revise, or retire—with evidence.

If your team is still deciding where to begin, start with Defining Your Real Estate AI Strategy and choose the first workflow before choosing more tools.

Five questions for the leadership team

1. Where are we on the peer maturity spectrum—honestly?
2. Which two or three workflows combine volume, pain, data access, and measurable value?
3. Who is the senior sponsor, and who owns the workflow every day?
4. Where must a human approve assumptions, exceptions, judgment, or external output?
5. What baseline would prove faster, cheaper, safer, or higher-capacity performance in 90 days?

Questions and caveats

What the next wave should test

• Which live workflows generate measured savings, revenue lift, faster decisions, or reduced risk after integration and maintenance costs?
• Which data and system constraints most often block underwriting, reporting, or property workflows?
• Which governance, permissioning, audit, and human-review standards are firms using today?
• Who becomes the durable operating owner after founder sponsorship creates initial momentum?

How to read this report

• This is a directional survey of 17 respondents, not a probability sample or market-sizing study; one response equals 5.9 percentage points.
• Results are self-reported.
• Multiple-selection questions sum to more than 100% by design, and cross-segment comparisons use very small denominators.
• Firm vignettes are anonymized reconstructions of individual response patterns, not direct quotations or independently verified case studies.