🤖 The 5 AI Prompts I'd Hand a New Property Manager on Day One

It's September. Pre-leasing for next June just opened, and you have 30 units to list by Friday. Half of them share a floor plan. You've written "spacious 2-bedroom close to campus" so many times it's lost all meaning.

You can spend the week staring at a blank listing, or you can spend an afternoon.

This isn't a list of AI tools. It's a list of jobs, with the exact prompt to paste and what to check before you trust the answer.

🔒 First, clean what you paste

Before uploading anything, remove names, phone numbers, email addresses, SSNs, bank information, dates of birth, signatures, and other identifying details. "Tenant A, Unit 4" is enough.

Check your account's data controls, and don't upload tenant or applicant records unless your company has approved that platform for business use.

🏷️ 1. Writing listing descriptions

Give it the facts and let it do the writing:

Write 3 rental listing descriptions for a 2-bed, 1-bath apartment. Details: $1,400/month, available June 1, in-unit laundry, off-street parking, 4 blocks from campus, updated kitchen, heat included. Under 100 words each. Describe the unit and location only, not the ideal tenant. Do not invent amenities or neighborhood claims.

Pick the best one, fix what's off, and post it. For units that share a floor plan, ask for variations so your listings don't all read the same.

The last two lines of that prompt matter. AI loves phrases like "perfect for young professionals" or "quiet adult building," which describe who should live there instead of what the unit offers. That's a fair housing problem. It will also happily add a pool you don't have. Read every listing before it goes live.

Learn more: Fair housing advertising guidelines, with examples of words to avoid

📄 2. Reading a document you didn't write

You picked up a new owner and inherited their leases. You need answers today, not after an afternoon of reading.

Summarize this lease. List term dates, rent and due date, late fees, renewal and notice terms, utility responsibilities, pet terms, and anything unusual. Give the section number for every item.

The section numbers are the trick. They let you check each answer in seconds. If it can't point to a section, don't believe it.

It can tell you what a clause says. It can't tell you whether that clause holds up in your state. That's still your attorney's call.

📊 3. Finding the pattern in your data

Export a report, remove tenant names, and upload it:

This is our work order log for the last 12 months. Which properties have repeat work orders for the same issue? Which categories are increasing? What should I look into first?

It will find patterns in minutes, like the same unit calling about the same drain every six weeks or a water heater brand that keeps failing.

It can spot patterns quickly, but every total, percentage, and property-level conclusion still needs verification. If a number is going to an owner, recalculate it yourself.

✍️ 4. The emails you keep rewriting

Late rent, noise complaints, parking violations, move-out reminders. You've written each of these a hundred times.

Draft a firm but respectful email to a tenant with three noise complaints this month. Reference the quiet hours clause. No threats. Under 120 words.

Once one comes out right, save the prompt. Next time, you only change the details.

🔧 5. Turning a text from your tech into a work order

"Bathroom fan not pulling, ceiling stained, tenant says it's been like that a while."

Turn this into a work order: location, observed issue, likely scope, what to check on arrival, and questions to ask the tenant.

Deciding what that stain means is still your job. The typing doesn't have to be.

Learn more: What that black mold work order actually means

🏠 The part that isn't optional

HUD issued guidance in 2024 on how fair housing law applies to AI and digital advertising. That guidance has been withdrawn and is no longer current agency policy. The Fair Housing Act itself hasn't changed, and neither has your liability.

Never let AI make an applicant-screening decision, alter your written criteria, or decide whether adverse action should be taken. Any screening or adverse action workflow needs approved criteria, documented human review, and legal guidance.

Learn more:

1. HUD's Fair Housing Act overview
2. The withdrawn 2024 HUD guidance on digital advertising, for background
3. NAA's summary of the federal fair housing rollback
4. Congressional Research Service legal overview of the Fair Housing Act

✅ AI drafts. A human signs.

If nobody on your team can say who signs, the process isn't ready.

🚀 Start smaller than the vendors want you to

Pick one job from this list. Only one. Run it for two weeks, and write down how long the task took before and how long it takes now.

The time savings are real, but they aren't the main benefit. A prompt that works every time is a documented process, and a documented process is something you can train a VA on. AI doesn't replace the person. It writes the playbook the person follows.

Every week, someone announces another AI-native platform for property managers. I've written about drowning in those demos, and I've publicly owned a software decision that didn't go the way we expected.

So when I tell you to start with one chat subscription and one repeatable prompt, that isn't resistance to new technology. It's experience. Prove that one workflow saves time before adding another platform to your stack.

Hit reply and tell me which job you picked. The best prompts will go in a future issue.