How to Use Claude AI for Commercial Real Estate- Part 3

Here’s a stat for you.

43% of CRE professionals already use tools like ChatGPT, Copilot, or Claude for basic tasks. That sounds like a great adoption number, until you read more: only one in six describe an actual CRE-native use case. The other five are doing what everyone does — writing, research, document summaries, basic editing. Separately, 44% of CRE firms report using AI chatbots at all.

But almost all of it happens in one place — a browser tab open to Claude.ai, typed into like it’s a Google search box, closed the second the answer comes back.

That’s the gap. Not adoption. Depth. Most people have Claude. Almost nobody is bringing it to the actual places they work. the property folder, the LOI they’re marking up, the comp sheet, the flier that marketing is behind on.

This post is a quick tour of six of those places with one realistic use case that I have built in couple of hours for my clients . With honest caveats where each one still has real limits. No hype, no “AI will replace your broker” nonsense. Just what it does well right now, and what it doesn’t.

WEB CHAT (Claude.ai)

Use case: You’re planning to attend the next ICSC event and want to think through how to connect with attendees ahead of time — messaging, timing, who to prioritize.

This is the one everyone already does, and it’s genuinely useful for it. Ask it to help you think through outreach sequencing, draft a few intro messages, or role-play how a cold LinkedIn note might land before you send it.

Limitation: it doesn’t know who’s actually attending, doesn’t have your contact list, and won’t remember this conversation next time you open a new chat unless you’re using memory or a project. Treat it as a thinking partner, not a directory.

You can also turn this conversation into a meta prompt. You give it a rough goal, and it builds a detailed, clean, and reusable template for each of these tasks that you can turn it into a skill.

DESKTOP

Use case: Use Cowork mode and point it at a property or portfolio folder with hundreds of leases and have it build a critical dates tracker — lease expirations, renewal windows, option deadlines — pulled straight from the documents already sitting there.

This is where it stops being a search box. No uploading files one at a time, no re-explaining what a rent roll is.

Limitation: The more the files the slower it will be so best to have Claude notify you when the task is done. It’s only as good as what’s in the folder. Inconsistent lease formats, scanned PDFs without OCR, or a messy naming convention will slow it down or cause it to miss something. Spot-check the first output before you trust the next twenty.

You can also have it build an artifact once you have reliable data. Claude Artifacts are dedicated, interactive windows that open alongside your chat to display self-contained content like dashboards. In the future you can simply ask it to refresh it or schedule refreshes.

DESIGN

Use case: A flier for a lease listing, or a first pass at an OM brochure — layout, copy, and visuals in one go from a one-line brief.

Limitation: it’s a first draft, not a design department. Brand consistency, print-ready specs, and the final polish pass still need a human eye. If you feed it prior examples and your actual brand assets ahead of time, you can get a great output.

And its expensive - three times more than using Cowork so be prepared to be on Max or Teams Premium plan if you plan to use it once a week or more.

WORD

Use case: LOI redlining by brokers or leasing reps. Feed it a playbook or your standard positions and it’ll mark up a lease or LOI with tracked changes and a rationale for each one.

This is one of the highest-leverage uses in the list. It turns a first-pass review from an hour into minutes.

Limitation: it redlines reliably only against the playbook you give it. It doesn’t have independent judgment on your risk tolerance, and it’s not a substitute for counsel on anything material. In the absence of a deterministic playbook and guidelines from you, it may miss a few significant clauses or enthusiastically redline every other line and the other side knows you just farmed the job to AI. Good first pass, not the final signature-ready draft.

EXCEL

Use case: A listing comparison for a tenant rep deal, or a net effective rent analyzer for a landlord rep deal — built with real formulas, not just a static table.

Where it earns its keep is speed on the model itself — formulas that update when an assumption changes, not a one-time calculation. You can also ask it to highlight its assumptions and the cells where it used them.

Limitation: garbage in, garbage out. It doesn’t know your actual market comps or your underwriting formula unless you give them to it, and any model it builds still deserves an audit pass before it goes in front of a client.

BROWSER

Use case: Troubleshooting a leasing application’s design, filling out a form automatically, watching a tutorial video and turning what it learns into a repeatable skill you can run again.

This is the most underused surface on the list, and arguably the most fun — it’s the one that turns a one-time task into something reusable.

Limitation: it can’t get past logins that need MFA it doesn’t have, can’t beat a CAPTCHA, and shouldn’t be left unsupervised on anything that submits or sends something on your behalf. Watch the first pass.

Six surfaces, six different jobs, one Claude. The people getting real value out of this aren’t the ones with the fanciest prompts. They are the ones who stopped treating it like a search engine and started bringing it to wherever the actual work is happening.

If you’re only using the chat window, you’re using maybe a sixth of what’s already sitting in your plan.