I quit SpaceX to become a leasing agent.
I know... ridiculous, right? One day I'm working on rockets that could theoretically take us to Mars, and the next I'm filling out paper rental applications in Redwood City, California for a small owner-operator who still requires money orders for rent.
But here's what I found that shocked me even more than the file cabinets and Excel-challenged site teams: I worked longer hours as a leasing agent than I did at SpaceX. The work was tedious, soul-crushing, and desperately in need of automation.
So we built Zuma. We built the tool I wished I had when I was on-site, something that could do the heavy lifting so humans could focus on what actually matters: building relationships and serving residents.
That was five years ago.
Fast forward to today, and nearly every company at AIM claims they're "AI-native." Every CRM suddenly has ChatGPT "plugged in." Every chat bot has been rebranded as an "AI leasing agent."
But here's what nobody's talking about: AI is making things worse, not better.
At this year's Apartment Innovation & Marketing Summit, we decided to do something different. Instead of another session celebrating AI's potential, we pulled back the curtain on what's actually happening when operators deploy AI without proper guardrails, clean data, or a clear strategy.
I even put on my old thespian hat (yes, I originally went to school for musical theater) and created a video showing what it's really like on-site when AI fails spectacularly. Triple-booked tours. Phantom amenities that don't exist. Residents asking about the "pet psychic on Tuesdays."
The audience's reaction? Immediate recognition. Hands shot up across the room when I asked who could relate to one or more of those AI fails.
This is the conversation nobody's having at industry conferences, but it's the one that matters most.
Joining me on stage were two of the strongest operators I know: Dalia Kalgreen from Unified Residential and Ryan Funt from Fitzrovia. Together, we dove into the fundamental disconnect that's breaking multifamily operations.
Here's what changed: Renters now expect instant gratification. They want answers at 2 AM. They want to book tours at 3 AM with the expectation of touring at 10 AM that same day. They want to track their maintenance request like they track their DoorDash order.
And you know what? That's not unreasonable. We've set those expectations in every other part of their consumer experience. But multifamily is stuck in the past.
Ryan shared some eye-opening research: Respond in 2 minutes and you'll see double the conversions. Wait longer than that? There's a linear drop-off between 2-10 minutes, and by the time you're at 60 minutes, prospects are 50% less likely to ever respond again.
Think about that. When I started coming to AIM five or six years ago, the goal was to respond within an hour. Then it was 45 minutes. Then 30. Then 20. Now it's 15 minutes and really, it should be 2.
Your leasing associates don't even have time for bathroom breaks anymore. That's insane. And that's exactly where AI should be filling the gap.
But here's the catch: 60% of prospects are engaging after hours. If you don't have a live "book a tour" form on your website in 2025—Class B, Class C, vintage, doesn't matter—you're officially operating in the stone ages.
Here's something fascinating that emerged in our discussion: renters are increasingly choosing to engage with AI over humans in the early stages of their journey.
Why?
It's not because AI is better at building relationships. It's because of Uncertainty Reduction Theory.
Renters are in research mode. They're comparing properties, de-risking their decision, and gathering information. They know humans are there to sell them. AI, on the other hand, gives them factual information without the sales pressure.
This creates an interesting paradox: good AI builds the trust that makes human interaction more valuable when it does happen.
AI should handle the simple, repetitive stuff: What's your pet policy? What are your rates? Are utilities included? It does a slam dunk job on these questions.
But when you get into big emotions, big feelings, complex negotiations, nuanced maintenance issues, that's when humans are irreplaceable. The key is nailing the handoff between the two.
Dalia hit this point hard: the biggest barrier isn't renters resisting technology. It's property management companies resisting change.
Some operators believe personalization can only come from humans. They expect renters to adapt to their processes instead of meeting renters where they are. And after COVID forced everyone to adopt technology, there are now people saying, "Let's throw it all out and go back to pen and paper because technology failed us."
No…bad implementation failed you. Technology that wasn't built natively for multifamily failed you.
Your Yardi, your Entrata, your RealPage, your MRI—these were built on infrastructure that isn't AI-first or AI-native. They're lagging behind, which is why you continue to see bugs and why connecting your tech stack feels like duct-taping solutions together.
And how many logins do your renters have to go through to get a lease at your properties? The audience responses ranged from 3 to 5 different platforms. That's a fragmented, poor experience.
Dalia asked a crucial question: Have you done a tech stack audit lately?
Most companies have deployed 20+ tools. They're spending money on overlapping solutions. Different departments are using different technologies without coordination. And nobody's doing the audit to figure out where the redundancies are or which tools actually move the needle.
When we asked this question a few months ago at another event, almost nobody raised their hand to say they'd done an audit recently.
Check out our Multifamily Budget Season Checklist to analyze your technology bloat.
If you're a VP, a director, a manager reading this: get everyone in a room—marketing, IT, maintenance, asset management—and do a tech stack audit together. You'll be shocked at what you find.
Only a couple hands went up when Ryan asked who's still using call centers. That's progress.
But for those still holding on: they're expensive, they have exponential turnover, they miss calls, they take sick days, and you're spending your time auditing calls instead of focusing on strategy.
AI can replace a call center in a day. Mostly accurate answers, 24/7 availability, zero sick days, no training required. It's a no-brainer in 2025.
If you're going to implement AI successfully—and I mean actually successfully, not just checking a box on a vendor pitch deck—you need to understand the Four C's:
You need documented playbooks, clean data, and corporate-to-site team alignment. AI can only learn from what's documented. If your processes aren't written down, if ownership isn't clear, if your data is messy then your AI will respond inaccurately.
Not every property is the same. A Class C asset in Houston needs a completely different approach than a Class A in Santa Monica. Your AI vendor needs to configure to your business needs, not force you to adapt to their one-size-fits-all model.
Most AI vendors have never worked on-site. They don't know what it's like. And you end up sacrificing the way you want to run your business for the way they think you should run it.
If your AI qualifies a prospect and then hands them off to a human who asks all the same questions again, well then you've failed. The experience needs to be seamless. The context needs to transfer. The conversation shouldn't restart.
Your AI shouldn't live in a separate portal from where your teams actually work. That's confusing and creates more work, not less.
It's easy to see a great AI demo. It's much harder to deploy that AI at scale across your entire portfolio and handle all the edge cases that come up.
Consistency means reliability. It means fair housing compliance. It means handling the thousand different ways humans show up and ask questions. This is where enterprise-grade AI separates itself from the noise.
Ryan and Dalia shared how they're actually using AI successfully in their operations.
Ryan at Fitzrovia: They can now SMS blast all active prospects in 30 seconds with a promotion. No downloading Excel sheets, no uploading to email platforms, no hunting for creative assets. Just done.
They're also using AI as a third data point to pressure-test their ad spend decisions—getting recommendations on where to pull back on Google Ads or push down on Meta.
Dalia at Unified Residential: They've automated rent collection reminders through "Kelsey" (our AI). At first, residents were confused—"Who is Kelsey and why is she telling me to pay rent?" But they quickly realized: Oh, this is helpful. She's trying to help me pay on time.
The result? They eliminated the assistant manager role entirely, centralized collections to corporate, and saw their collection rates shoot through the roof in tough markets. All by automating with a human-first implementation.
I learned this lesson the hard way in Redwood City. The property I worked at was working-class. Most residents didn't have computers. Many didn't have email addresses. But everyone—everyone—had a cell phone.
That's why we built Zuma starting with SMS. Not because it was trendy. Because it was the only way to actually reach people where they were.
That principle extends beyond just the technology channel. It's about understanding who your renters are, who your team members are, and designing solutions that actually work for them, not solutions that force them to adapt to you.
And here's the uncomfortable truth: if you're in corporate and you haven't been on-site in the last week, if you haven't answered the phone, if you don't know what your site teams are dealing with on a daily basis—you're making decisions in a vacuum.
The pendulum swung hard toward technology. Now it's time to bring it back to center.
The future isn't AI replacing humans. It's AI enabling humans to be more human.
Your onsite teams shouldn't be spending their time on repetitive data entry, chasing rent checks, or answering "What's your pet policy?" for the hundredth time. They should be building relationships, solving complex problems, and creating community.
AI should be the tool that frees them to do what only humans can do: show empathy, read nuance, make judgment calls, build trust.
But that only works if you implement AI correctly. With clean data. With clear playbooks. With proper handoffs. With the right vendor who understands multifamily operations because they've actually lived them.
Look, I know AI is the buzzword. I know every vendor is claiming they're "AI-native." I know it's tempting to check the box and say you've implemented AI.
But if your AI is creating chaos instead of clarity, if it's frustrating prospects instead of serving them, if your site teams are working around it instead of with it—you haven't implemented AI. You've just added another tool to the pile.
The question isn't whether AI exists in your tech stack. The question is: Is it built natively for multifamily? Does it connect into your current operations? Is it actually making things better?
Because if the answer is no, you're not alone. Most operators are dealing with this. But unlike the vendors on stage pitching their "solutions," we're the ones willing to talk about it honestly.
That's what we did at AIM. That's what we'll keep doing at Zuma.
We're building AI the way it should have been built from the start: by people who've actually worked onsite, for people who are running operations every day, with the humility to know that technology is a tool—not a replacement for human judgment.
And if you've experienced AI fails in your portfolio, you're not alone. Let's fix this together.
with your current PMS and CRM
