Anyone using LLMs (Claude, OpenAI, Grok, etc) to manage their rentals in any way?
Patrick KrawczykowskiPro Member
Investor · Chicago, Chicago IL · Member since 2026 · 46 posts · 55 votes
Anyone using LLMs (Claude, OpenAI, Grok, etc) to manage their rentals in any way?
Most Popular Reply
Honolulu, HI · Member since 2008 · 3k+ posts · 2k+ votes
3moAlthough retired from PM, I have been learning and using Claude Coworker and Code on some small projects, although I mostly use Grok for minor lookups and small tasks. Based on a lot of comments on this site, it would be nice if there was a place to share "prompts", "skills", and "agents" as actual examples and for learning. There seems to be a lot of either really uninformed efforts, or pro IT guy level praise...usually followed by a vague, or not so vague, sales pitch. I'm not from Missouri, but Show Me!
17 Replies
- Honolulu, HI · Member since 2008 · 3k+ posts · 2k+ votes3mo
Although retired from PM, I have been learning and using Claude Coworker and Code on some small projects, although I mostly use Grok for minor lookups and small tasks. Based on a lot of comments on this site, it would be nice if there was a place to share "prompts", "skills", and "agents" as actual examples and for learning. There seems to be a lot of either really uninformed efforts, or pro IT guy level praise...usually followed by a vague, or not so vague, sales pitch. I'm not from Missouri, but Show Me!
- Investor · Chicago, Chicago IL · Member since 2026 · 46 posts · 55 votes3mo
Quote from @Richard F.:
Although retired from PM, I have been learning and using Claude Coworker and Code on some small projects, although I mostly use Grok for minor lookups and small tasks. Based on a lot of comments on this site, it would be nice if there was a place to share "prompts", "skills", and "agents" as actual examples and for learning. There seems to be a lot of either really uninformed efforts, or pro IT guy level praise...usually followed by a vague, or not so vague, sales pitch. I'm not from Missouri, but Show Me!
I love the idea about sharing skills and prompts to the community to level up all the members, let's see what we can do there.
- 3moRachid AbadliBusiness MemberInvestor · Sacramento, CA · Member since 2024 · 129 posts · 37 votes
Great question, @Patrick Krawczykowski. I use AI every day, but not as a chatbot.
The biggest value has been compliance and workflow automation. Instead of spending 30 minutes figuring out whether I can increase rent on a specific unit, I enter the address and lease details and get the applicable rent cap, notice period, and any local rules. I use it the same way for notice drafting and maintenance documentation.
The biggest lesson for me is that landlords don't want another chat window. They want AI embedded in the work they're already doing. If it saves time without changing their workflow, they'll use it. If it feels like "talking to AI," most won't.
LeaseBase.ai- Investor · Chicago, Chicago IL · Member since 2026 · 46 posts · 55 votes3mo
Quote from @Rachid Abadli:
Great question, @Patrick Krawczykowski. I use AI every day, but not as a chatbot.
The biggest value has been compliance and workflow automation. Instead of spending 30 minutes figuring out whether I can increase rent on a specific unit, I enter the address and lease details and get the applicable rent cap, notice period, and any local rules. I use it the same way for notice drafting and maintenance documentation.
The biggest lesson for me is that landlords don't want another chat window. They want AI embedded in the work they're already doing. If it saves time without changing their workflow, they'll use it. If it feels like "talking to AI," most won't.
Thanks for the feedback. Hear you on the "not another chatbot" request. We're not wanting to add complexity to anyone's workflow. But for deal and rent analysis, we're wanting to make sure we support those AI native workflows.
- Specialist · Goa, India · Member since 2026 · 175 posts · 36 votes3mo
Completely agree with Rachid's point — that's the pattern I've seen most consistently too. Self-managing landlords and small operators don't want a new interface, they want the exhausting parts of their existing one handled quietly in the background.
One example that's worked well: instead of a chatbot tenants have to go talk to, using AI to triage the messages that already come in through the channels tenants use (mostly texts) — sorting the "can wait till business hours" stuff from the "2am, actual emergency" stuff, and pre-drafting responses for the repetitive 90% so the landlord isn't typing the same answer for the fifth time that week. The landlord still decides, they're just not the one doing the sorting anymore.
Patrick, since you're gathering feedback — curious whether BP's product thinking is leaning toward a chat-style AI surface, or more toward this kind of "quietly embedded in what people already do" direction? Feels like the second is where the actual time-savings are, based on what I hear from landlords, but genuinely curious what you're seeing from the product side.
- Investor · Chicago, Chicago IL · Member since 2026 · 46 posts · 55 votes3mo
Quote from @Andrea Fernandes:
Completely agree with Rachid's point — that's the pattern I've seen most consistently too. Self-managing landlords and small operators don't want a new interface, they want the exhausting parts of their existing one handled quietly in the background.
One example that's worked well: instead of a chatbot tenants have to go talk to, using AI to triage the messages that already come in through the channels tenants use (mostly texts) — sorting the "can wait till business hours" stuff from the "2am, actual emergency" stuff, and pre-drafting responses for the repetitive 90% so the landlord isn't typing the same answer for the fifth time that week. The landlord still decides, they're just not the one doing the sorting anymore.
Patrick, since you're gathering feedback — curious whether BP's product thinking is leaning toward a chat-style AI surface, or more toward this kind of "quietly embedded in what people already do" direction? Feels like the second is where the actual time-savings are, based on what I hear from landlords, but genuinely curious what you're seeing from the product side.
We’ll likely try many approaches to helping more investors connect and grow using LLMs, but I agree that embedded tools in workflows that investors already use seems to be the win win scenario that we’ll aim for.
- Real Estate Agent · Memphis · Member since 2026 · 626 posts · 360 votes3mo
I'm finding AI is most useful for the administrative side of property management rather than replacing the decision-making.
Things like drafting tenant communications, summarizing maintenance requests, organizing lease or inspection notes, creating SOPs, and helping with documentation are all good use cases. It saves time, but I still want a human reviewing anything that's tenant-facing or has legal implications.
I also think there's a lot of potential in using AI to surface patterns—like recurring maintenance issues, lease expirations, or delinquency trends—so owners can be proactive instead of just reacting. That's where I think it starts becoming a real operational tool rather than just a writing assistant.
- Member since 2025 · 7 posts · 1 vote3mo
Yeah — this is basically what I do all day. The useful pattern isn't "AI runs my rentals," it's AI handling the repetitive language + follow-up work: drafting renewal notices, triaging maintenance texts into categories, chasing missing vendor COIs, turning a messy owner question into the right report. The judgment stays with you; the typing and the nagging gets automated. Happy to share what's actually held up vs. what's hype.
- Investor · Get yourself trained before doing something inadvisable. · Member since 2024 · 3k+ posts · 1k+ votes3moOne of my students ran into an issue he was having at one of new properties regarding a purchase. And being the clever guy that he is, turned to AI to answer the issue. He gave it to me to take a look at. It looked good from the outside for a newer investor, AI was even "polite" but the AI missed something very important that would have gotten my student sued. I don't care for that to happen so we rejected it and followed the law. It isn't what you know that gets you into trouble, it's what you don't know.
Quote from @Patrick Krawczykowski:
Pretty much title.
I wonder if anyone is using AI in a meaningful way? Renter communication? Maintenance reminders/upkeep/scheduling? Something else?
Disclaimer, I just joined bigger pockets as the Head of Product. I'll likely be looking for feedback from this community on here quite a bit. Appreciate any and all response.
- Patrick
- Investor · Chicago, Chicago IL · Member since 2026 · 46 posts · 55 votes3mo
Quote from @Ken M.:
One of my students ran into an issue he was having at one of new properties regarding a purchase. And being the clever guy that he is, turned to AI to answer the issue. He gave it to me to take a look at. It looked good from the outside for a newer investor, AI was even "polite" but the AI missed something very important that would have gotten my student sued. I don't care for that to happen so we rejected it and followed the law. It isn't what you know that gets you into trouble, it's what you don't know.Quote from @Patrick Krawczykowski:
Pretty much title.
I wonder if anyone is using AI in a meaningful way? Renter communication? Maintenance reminders/upkeep/scheduling? Something else?
Disclaimer, I just joined bigger pockets as the Head of Product. I'll likely be looking for feedback from this community on here quite a bit. Appreciate any and all response.
- Patrick
Interesting, can you share more details about what was missed specifically?
- Investor · Get yourself trained before doing something inadvisable. · Member since 2024 · 3k+ posts · 1k+ votes3moIt was an issue regarding title in Arizona. Since I teach how to control properties using creative finance, he was wanting to use the concept and he trying to do a purchase. He was getting hung up at title & escrow. The AI wasn't able to distinguish between Subject To and an executory contract and the answer it gave assumed they were both the same thing. They are not. I walked him through the conflict and he bought the property, but title was very confused about the mixed messages coming from AI. It returned a warning at the bottom stating "ChatGPT is Al and can make mistakes.". No kidding? Now THAT'S truth in advertising.
Quote from @Patrick Krawczykowski:
Quote from @Ken M.:
One of my students ran into an issue he was having at one of new properties regarding a purchase. And being the clever guy that he is, turned to AI to answer the issue. He gave it to me to take a look at. It looked good from the outside for a newer investor, AI was even "polite" but the AI missed something very important that would have gotten my student sued. I don't care for that to happen so we rejected it and followed the law. It isn't what you know that gets you into trouble, it's what you don't know.Quote from @Patrick Krawczykowski:
Pretty much title.
I wonder if anyone is using AI in a meaningful way? Renter communication? Maintenance reminders/upkeep/scheduling? Something else?
Disclaimer, I just joined bigger pockets as the Head of Product. I'll likely be looking for feedback from this community on here quite a bit. Appreciate any and all response.
- Patrick
Interesting, can you share more details about what was missed specifically?
- Specialist · Goa, India · Member since 2026 · 175 posts · 36 votes3mo
@Patrick Krawczykowski Good to hear, especially coming straight from the product side. Excited to see where BP takes it.
- Specialist · Goa, India · Member since 2026 · 175 posts · 36 votes3mo
@Patrick Krawczykowski Good to hear, especially coming straight from the product side. Excited to see where BP takes it.
@Sonia Trevino This matches almost exactly what I've seen too — it's the categorizing and chasing that eats the time, not the judgment calls. What's been your biggest "sounded good in theory, didn't hold up" example? - Property Manager · Scottsdale, AZ · Member since 2016 · 82 posts · 13 votes2w
Yes, every day. Full disclosure: I asked my AI chief of staff to sum this up, because I'd never get around to writing it myself. Kind of proves the point.
I'm a broker-owner in Scottsdale: buy, sell, leasing placement and property management, plus my own rentals. AI runs a big share of the back office across all of it. What I've learned:
1. AI is not the system of record. Every deal, tenant, lease, mortgage and vendor lives in one place (I use Jira; a PM platform or even Airtable works). The AI reads and updates it. Everything it learns about how I work gets written to a wiki, so any other AI tool I bring in reads that first and is up to speed fast. If the chat is your only memory, it drifts and you won't notice.
2. Rules first, AI second. Anything predictable runs on plain automation. AI takes the messy work: reading documents, drafting, comparing.
3. Where it actually pays off:
Buy side
Contract to close: it reads the executed contract and addenda, sets every deadline (inspection, BINSR, seller response, buyer election, COE) and flags what's coming. I verify every date. A missed inspection deadline can put your client's earnest money at risk, so that one gets its own alarm
Out-of-state buyers: I narrate a video tour, it turns the transcript into a client-ready summary
Inspection report into a short decision brief: what matters, what's cosmetic, what to ask for
Sell side
Listing intake into a launch brief, and a check that the facts match across MLS, flyer and listing remarks
Leasing placement and prospecting
An MLS feed of new and canceled rentals every morning, screened against my rules: agent-owners, PM companies and furnished seasonal listings drop out automatically
An owner check from county records before I dial, then a call queue with notes. Every "no" becomes a rule so the same dead lead never comes back
Property management
Vendor invoices land in the work order automatically and get read against the photos and the job, so the owner sees it all lined up before paying the vendor directly
One dashboard: HOAs, insurance and lease expirations, renewals, open repairs
Contractor and insurance quotes normalized side by side, gaps flagged
My own rentals
Rent and mortgage funding: who paid, which account is short before the draft, cash flow and equity per door
4. It's confidently wrong sometimes. My dashboard once called a BINSR late that I'd delivered the day before, because it read the ticket status instead of my notes. Make it show where every number comes from.
5. Protect client data. Turn off model training, keep SSNs and account numbers out of chats.
Start with one painful recurring task, not the whole operation. What's eating the most of your time right now?
- Investor · Nashville, TN · Member since 2026 · 6 posts · 1 vote2w
@Richard F. asked for actual examples instead of pitches, so here's one you can copy.
I ran manufactured-housing communities for about ten years, and maintenance triage was one of the back-office systems I built. This is the version I'd set up today. Paste it into a ChatGPT Project, a Claude Project or a Gemini Gem as the instructions, then paste tenant texts in as they arrive:
---
You are the maintenance intake coordinator for [community name], a [#]-lot community in [city, state].
For every tenant message I paste in:
1. Sort it as EMERGENCY (active water leak, no heat below 40°F, gas smell, sewage backup, sparking or burning smell), URGENT (needs a response within 24 hours) or ROUTINE.
2. Pull out the lot number, whether the home is tenant-owned or park-owned, what's broken and since when.
3. List what the tech still needs to know before rolling a truck.
4. Draft a reply to the tenant in under 40 words. Plain language. Never promise a time or a cost.Rules:
- Gas smell: the reply tells them to leave the home and call the gas company first. No exceptions.
- Tenant-owned home with a problem inside the home: flag it for me as "possible tenant responsibility, check lease section [X]." Don't say that to the tenant. I decide.
- If you're unsure of the category, pick the more urgent one and tell me why.---
Two things make it work. First, write it like you're briefing a new hire on day one. If you can't write the brief, the process isn't ready for AI or for a person. Second, keep the judgment calls with a human: who's responsible, anything involving money, anything with legal weight. The AI handles the sorting and the first draft.
What kind of message do you answer most often? I'm happy to sketch a brief for it here in the thread.
- Technology · Santa Cruz, CA · Member since 2025 · 15 posts · 0 votes1w
Alik's rule is the one that held up for us too. The AI is not the system of record, it reads from one.
Since Richard asked for real examples, here is the maintenance side of that. Our customers' techs clock in and out of jobs on their phone, GPS keeps the time on site and the miles between properties, and that record lives in Orca. Then they connect Claude to it the same way you add any other connector and ask things like "how many hours did we put into 123 Main in October" or "which tech had the most drive time last week." Nobody retypes anything into a chat window. The AI is just a faster way to ask the data a question, and the answer traces back to a clock-in with a GPS stamp.
I built Orca, so grain of salt, but that part is running today, not a roadmap slide.
Join the conversationCreate a free account to reply, vote on answers and follow this thread.