AI Cold Email for CRE Brokers: How to Use It Without Sounding Like AI

There's a real problem with AI cold email in commercial real estate right now. The tools have gotten good enough that brokers are using them at scale, and the recipients (landlords, tenants, owners) can tell. Generic AI emails are landing in spam folders, getting deleted on sight, and burning sender reputation in ways that hurt the brokers' actual deals when they need to reach people for real.

But used well, AI is a genuine accelerator for cold outreach. Drafting time drops from 20 minutes to 90 seconds per email. The trick is using it in a way that doesn't sound like AI, and that respects the recipient's time and intelligence.

AI cold email works for commercial real estate brokers when the broker brings the signal and the relationship context, and uses the AI to accelerate the drafting and personalization layer. It fails when brokers use AI to scale generic outreach, because landlords, tenants, and owners have learned to identify and ignore obvious AI output. The mechanical rules: every email needs a specific reason for the contact (a recent closing, ownership change, expansion announcement, mutual connection), the AI draft needs human editing to remove the AI tells (excessive structure, formal tone, em dashes, banned phrases like "I hope this finds you well"), the volume per recipient should stay low (one to three touches before moving on), and the broker has to be ready for the response with real follow-through. The right tools combination for most brokers: a signal source for the hook (Station CRM, trade press, ACRIS), an AI drafting tool (ChatGPT, Claude, or a CRM with built-in AI), a real sending address (not a marketing automation domain), and a follow-up workflow tracked in a CRM. Station CRM combines all of this in one place. Solo brokers using ChatGPT plus Gmail can run the same play with more manual effort.

Why most AI cold email in CRE is failing

A few specific reasons:

The recipients have pattern-matched on AI output. Landlords and tenants have been receiving AI-drafted emails for over a year. They know the tells. Em dashes in the wrong places. "I hope this email finds you well." Vague references to "exciting opportunities." Three-sentence paragraphs with a question at the end. The moment a recipient identifies the email as AI-drafted, they assume zero personal effort went into it and they delete.

Generic outreach has no reason for the contact. Most AI-drafted CRE cold emails are missing the specific signal that justifies the outreach. Without "I noticed your tenant at [address] announced their closing yesterday," the email is just "I'd love to discuss your portfolio." That email isn't worth the recipient's time, and they know it.

Volume is treated as a solution. Brokers using AI to send 200 emails a day are signaling to inbox providers that they're a spam source. The deliverability suffers, the domain reputation suffers, and the broker's real emails to their real contacts start landing in spam too.

The follow-up isn't there. A response to a cold email that doesn't get a thoughtful follow-up within a day kills the relationship before it starts. AI can draft the initial email. The follow-through has to be real.

What the broker has to bring

AI cannot supply the things that make cold outreach work. The broker has to bring:

The signal. A specific, recent, identifiable event that justifies the outreach. A retail tenant closing. An ownership change on a building. A brand announcing expansion. A lease expiration coming up. A 1031 exchange seller window. The signal is the reason the recipient reads the email instead of deleting it.

The relationship context. Have you done a deal in this corridor before? Do you know someone the recipient knows? Are you the broker for a competing space? Are you a tenant rep with a real client looking for this kind of space? Context that the recipient can verify or recognize gets the email taken seriously.

The specific ask. A 10-minute call. A meeting next week. A response with their thoughts on a specific question. Generic asks ("let me know if you'd like to discuss") get ignored. Specific asks get responses.

The judgment about who to contact and when. AI can produce 100 emails. The judgment about which 10 of those should actually go out, and when, is the broker's.

What AI can do well

Given the broker brings the above, AI can accelerate:

The drafting itself. A well-prompted AI gets you a 4-sentence email in 30 seconds. You edit it down to a 3-sentence email in another 30 seconds. Total time per email: about 90 seconds for a clean, specific draft.

The personalization layer at modest scale. You have a list of 20 landlords you want to reach about a corridor signal. You give the AI a brief on each (who they are, what they own, the specific angle for them) and have it draft individualized emails. You spend 3 minutes reviewing each. Total time: 90 minutes for 20 high-quality emails. That's a 3x to 5x speedup over writing them all by hand.

Voice consistency. If you give the AI examples of your past emails, it can match your voice on new drafts. This is genuinely useful for brokers who want their outreach to read consistently.

Subject line variations. Quick A/B testing on subject lines. The AI can generate 5 options in 10 seconds. You pick the one that reads least AI-generated.

A prompt that produces useful output

The prompt template that works:

"Write a 3-sentence cold email to [name], who is [role at company]. The reason I'm reaching out: [specific signal, with date and source if possible]. My ask: [specific ask with timeframe]. About me: I'm [role] at [brokerage], I've done [relevant credibility, e.g., 'three deals in this corridor in the past 18 months']. Voice: direct, no greetings like 'I hope this finds you well,' no exclamation points, no em dashes. Write it the way a working broker would, not the way a marketer would."

The output will need editing. Take out the AI tells. Tighten the language. Make sure your specific voice comes through. Then send.

Things to scrub from every AI draft

Specific things to delete or rewrite before sending:

Em dashes. The AI loves them. Replace with commas, periods, or parentheses.

"I hope this email finds you well." Delete on sight. No working broker writes that anymore.

"I wanted to reach out to discuss." Replace with the actual reason.

"I'd love to learn more about your portfolio." Delete. Replace with the specific question.

"Exciting opportunity," "tremendous potential," "synergies." Delete.

Three-sentence paragraphs with a forced question at the end. Restructure to flow naturally.

Capitalized headers in a short email. Delete. Short emails don't need headers.

Overly formal salutations. "Dear Mr. Smith" reads as AI. "Hi John" reads as human.

The volume question

How many AI-accelerated cold emails per day is reasonable for a CRE broker?

The honest answer: it depends on your signal density. If you have 5 high-quality signals on a given day (real closings, ownership changes, expansion announcements), 5 to 15 outreach emails based on those signals is fine. If you have 50 signals, more is reasonable. If you have one signal and you're padding the volume with generic outreach, you're hurting yourself.

The math that works in 2026: low volume, high quality, real signals, real personalization, real follow-through. The math that doesn't work: high volume, generic content, scraped lists, no follow-up.

For perspective: a focused broker doing 10 to 25 outreach emails per day off real signals produces meaningfully more deals than a broker sending 200 a day of AI-generated generic content. The economics flipped on volume-based outreach a few years ago. They have not flipped back.

How Station CRM thinks about this

Station CRM was built around this specific problem. The morning briefing surfaces the signals (closings, sublease postings, ownership transactions, brand expansions). The AI chief of staff drafts outreach in your voice, using the signal as the hook and the recipient context as the personalization. You review and send from your real inbox. The follow-up sequence lives in the CRM so nothing falls through.

The point is not to scale generic outreach. The point is to compress the time from "signal happens" to "appropriate broker outreach in front of the right person" from hours to minutes, while keeping the broker in the loop on every send.

The honest take

AI is going to keep getting better at writing cold email. The recipients are going to keep getting better at filtering it. The competitive advantage doesn't lie in the AI itself. It lies in the broker's ability to surface signals others don't see, write outreach that connects to those signals in a way the recipient finds worth responding to, and follow through on the responses in a way that builds the relationship.

The brokers who use AI as a force multiplier on real signals and real judgment win. The brokers who use AI to scale generic outreach lose, slowly at first and then quickly.

Station CRM combines daily signal-based prospecting, AI outreach drafting in your voice, and the follow-up workflow in one place. Request a demo to see what an AI-assisted CRE workflow looks like when the signal layer is real.