How-To Guide

How to Use AI for Commercial Real Estate in 2026: Leasing, Investment Sales, Underwriting & Market Reports

Published May 2, 2026 · 14 min read

TL;DR

AI compresses broker-analyst work: comps, stacking plans, market reports, underwriting, LOIs, OMs.
Ten prompts below cover tenant rep, landlord leasing, investment sales, debt, BOVs, and quarterly reports.
Confidential data (rent rolls, T-12s, tenant names) stays in enterprise tooling with DPAs — never consumer ChatGPT.
RESPA, state licensing, ADA, and SEC Rule 10b-5 all apply to AI-drafted marketing and OMs.
The deal narrative, tour, and negotiation are still the broker's — that's where the fee lives.

Where AI fits in a 2026 CRE brokerage

A typical mid-market CRE broker runs 8-15 active listings or assignments per quarter, fields 40-80 tenant reps in a decent market, and produces 12-20 tour packages, 20-40 LOIs, and 4-6 BOVs or OMs per year. The analyst layer used to do the scaffolding — stacking plans, comps pulls, market slides. In 2026 AI does most of that layer in minutes.

What's left for the broker is the deal: the relationship, the tour, the negotiation, the judgment on which tenant the landlord will actually take, and the narrative that convinces an institutional buyer. CBRE's AI platform, JLL GPT, Colliers' and Cushman's tooling — and the 2026 wave of Happycapy, Claude, and Copilot deployments — all point the same direction: analytical scale is now a commodity. Differentiated judgment is the moat.

The 2026 CRE AI stack

Layer	Tool	Use
Comps & market	CoStar AI, CompStak AI, VTS, Crexi AI	Lease & sale comps, absorption, trends
Ownership & debt	Reonomy, Lev Capital AI, Real Capital Analytics	Owner outreach, debt sourcing, transaction history
Portfolio / property	Cherre, Yardi AI, MRI	Portfolio analytics, T-12 normalization
Marketing	Buildout AI, Apto, Biproxi, Rofo	OM builds, tour books, flyers, listing distribution
Writing & ops	Happycapy Pro, Claude for Work, Microsoft 365 Copilot	LOIs, market reports, BOVs, emails, QBRs

Ten copy-paste prompts for a 2026 CRE broker

All prompts assume enterprise tooling and senior-broker QA before anything leaves the shop. Replace bracketed content with your specifics and verify numbers against source systems.

1. Tenant-rep requirements brief from kickoff call

You are a CRE analyst. Input: tenant-rep kickoff transcript [paste]. Produce a structured requirements brief: company, use, size range, geography, must-haves (timing, parking, power, loading, signage, floor-plate), nice-to-haves, budget ($/SF base + OpEx), term, TI expectations, growth path, decision-maker, timeline. Flag anything that sounds like a protected-class proxy or a fair-housing-adjacent comment.

2. Submarket tour shortlist with stacking-plan hits

Based on the tenant brief [paste] and our CoStar/VTS pull of available blocks [paste CSV], produce a ranked tour shortlist of 8-10 buildings. For each: address, landlord, block size, floor, asking rate, OpEx, parking ratio, recent comps in-building, known lease expirations on floors above/below, signage availability, and two-sentence fit rationale. Flag anything over 15% above market so we don't waste a tour.

3. LOI draft (business terms only — counsel reviews legal)

Draft a non-binding LOI for [tenant] at [building, suite]. Inputs: the tenant brief, the landlord's published terms, and our target outcome [paste]. Cover: premises, term, commencement, rent (base + escalations), OpEx treatment, free rent, TI allowance, parking, signage, expansion/ROFO, renewal, security deposit, broker commission, assignment/sublet, early termination, and AS-IS condition. Mark DRAFT — ATTORNEY REVIEW REQUIRED. Do not include legal indemnification language.

4. Landlord leasing: stacking plan + retention strategy

For [building] inputs: current rent roll [paste anonymized], upcoming expirations next 24 months, market rate, in-place rates, and tenant industries. Produce: (a) stacking plan with color-coded expirations, (b) retention priority list (keep at-cost vs re-tenant at market), (c) renewal-offer outline per tenant with rate, TI, and term, (d) backfill scenarios for any block >10,000 SF.

5. Investment sales underwriting (broker BOV scaffold)

Underwriting scaffold for [property]: inputs — T-12 operating [paste], rent roll [paste], market OpEx benchmark, comparable sales (last 24 months in submarket), debt market read. Output: normalized NOI, stabilized NOI, vacancy and credit-loss adjustments, TI/LC reserve, cap-rate range with three scenarios (A-, B+, B), value range, implied $/SF, IRR to buyer at 65% LTV. Mark EVERY NUMBER that needs broker verification. Do not populate without source data.

6. Offering memorandum narrative (institutional-grade)

Draft the OM narrative sections for [property]. Inputs: underwriting, tenant roster, submarket read, property description, capex history. Sections: investment highlights (5-7), property overview, tenancy, submarket, demand drivers, value-add thesis, financial summary pointer. Tone: institutional, numerate, specific. No marketing-speak ("trophy," "irreplaceable," "prime") unless backed by a comp. Every factual claim needs a footnote placeholder for source.

7. Quarterly submarket report

Produce the Q[X] [YYYY] submarket report for [submarket, product type]. Inputs: CoStar pull [paste], CompStak leases [paste], notable sales, construction pipeline, tenant-in-market list. Sections: headline (3 bullets), vacancy/absorption, rent trend, concession trend, deliveries/pipeline, leases of note, sales of note, outlook. Include one chart description for a designer. Keep factual and quote-worthy — this is BD content.

8. Cold-outreach to property owner (Reonomy match)

Draft a cold outreach to the principal at [LLC/owner entity] re: [property]. Inputs: Reonomy snapshot (last sale, debt maturity, ownership tenure), current market comp for the asset, our relevant sale. Tone: peer, specific to their portfolio, under 120 words, clear ask (15-min call, not a pitch). Do NOT include "I noticed your property" or any generic pattern. Include one specific, numerate market observation.

9. Tour book / flyer copy

Write the copy for [building's] tour book. Inputs: building specs, amenity set, tenant roster (only publishable ones), transit, submarket. Sections: hero statement (20 words), building story (100 words), specs grid, amenities, neighborhood, sustainability (LEED/ENERGY STAR/WELL). No hyperbole. Match the building class — a B+ asset shouldn't read like a trophy tower.

10. Broker QBR for principal client

Draft the QBR for [client portfolio/client tenant rep]. Inputs: activity (tours, LOIs executed, leases, renewals, dispositions), market read, pipeline, fee/comp summary. Sections: scorecard, wins, open pursuits, market signal, decisions we need from you this quarter. One page. Tone: partner, candid, numerate — not vendor. End with three specific decisions requested.

Common mistakes to avoid

Shipping OMs without number verification. AI hallucinates cap rates and SF. Every number gets a broker signoff against source data.
Using consumer ChatGPT with client confidential info. Rent rolls and T-12s are confidential. Enterprise tenant with DPA only.
Generic outreach at scale. Principals recognize AI-pattern cold emails instantly. Specificity and numerate observation are the new moat.
Hyperbole in marketing copy. "Irreplaceable trophy" without comps backs up gets laughed out of institutional IC. Match claim to evidence.
Forgetting state licensing and RESPA. AI-generated referrals and fee-sharing language still subject to federal/state rules.

A 60-day rollout that respects CRE's quirks

Weeks 1-2: Enterprise LLM tenant with DPA. Compliance, IT, and GC sign off on data classes (public market vs client confidential vs regulated PII).
Weeks 3-4: Pilot on market reports and tenant-rep briefs (lower risk). Measure time-to-draft and senior-broker edit rate.
Weeks 5-6: Extend to LOIs and BOVs. Every LOI gets attorney review; every BOV gets senior-broker number verification with signoff sheet.
Weeks 7-8: Roll out OM and QBR workflows with principal-grade copy guidelines. Create a review checklist for hyperbole and source footnotes.
Ongoing: Quarterly tool review, semi-annual CLE / state-license update on AI rules, annual cybersecurity review for any tool touching tenant PII.

Frequently Asked Questions

Will AI replace commercial real estate brokers?

No, but it will compress what first-year associates used to do — comps research, stacking plans, market reports, LOI drafts. Brokers who adapt keep the client, the tour, and the negotiation. Those who don't cede analytical work to AI-native competitors. CBRE, JLL, Colliers, and Cushman have all shipped internal AI tooling in 2025-2026 — if you're a boutique broker, your AI stack is now competitive, not optional.

Can I paste client rent rolls into ChatGPT?

Not in consumer plans. Rent rolls contain tenant names, lease rates, expiration dates, and sometimes personal guarantees — confidential under most listing and representation agreements. Use enterprise tooling with a DPA: your brokerage's enterprise tenant of Happycapy Pro, Claude for Work, or Microsoft 365 Copilot, or a purpose-built CRE AI like CompStak AI, Reonomy, Lev Capital AI, or Cherre.

Is AI-generated offering memorandum content legal?

Yes, but it must be accurate. Misrepresentation in an OM is actionable under state real estate law and, for syndicated deals, SEC Rule 10b-5. AI hallucinates comparables and cap rates — verify every number against your source data, your CRM, CoStar/CompStak, county assessor, and the actual rent roll before distribution. Document your QA process in case of later dispute.

Which AI tools actually work in a 2026 CRE shop?

Minimum viable: CoStar AI (comps + market), CompStak AI (lease comps), Reonomy (owners), your brokerage's enterprise LLM (Happycapy Pro/Claude for Work/Copilot), and Buildout or Apto AI for marketing automation. Nice-to-have: Lev Capital AI for debt sourcing, Cherre for portfolio analytics, and Biproxi for smaller investment sales listings. Avoid anything that charges per query without a DPA.

What's the biggest mistake CRE brokers make with AI right now?

Generating generic OMs and market reports that read like every other broker's output. In a market that rewards differentiated information, AI-flattened content is a negative signal to institutional buyers and credit tenants. The fix is to use AI for the 80% scaffolding and reserve the last 20% (thesis, deal narrative, tenant rationale) for senior human judgment. That's where your fee comes from.

Want one workspace for OMs, BOVs, LOIs, and market reports?

Happycapy Pro runs tenant-isolated on an enterprise plan with a DPA, supports the full CRE writing stack (narrative, underwriting scaffolds, QBRs), and keeps client data out of consumer models. 50+ skills for spreadsheet analysis on T-12s and rent rolls, deck drafting for tour books, and research synthesis for submarket reports.

Try Happycapy Pro →

← Back to all articles