Outcome-First AI: Kyle Erb’s Vision of the Future

🎧 Listen to the Audio Version

Property managers don’t need more hype about AI. They need clarity.

That’s what I found in a recent conversation with Kyle Erb, Director of Systems at Real Property Management.

Kyle isn’t just theory-smart. He has deep domain expertise in proptech, and rare visibility across RPM’s hundreds of franchisees.

He knows what’s possible today.

He’s seen where the tools fall short. And he has a clear point of view on how AI will actually transform this industry.

More importantly, Kyle helped me make sense of ideas I had been wrestling with for months. He made them click.

Here are six lessons I took from our conversation on the Property Management Frame Breakers podcast.

What Kyle Knows About AI That Most Don’t Understand (Yet)

Kyle has tested the tools. He has run the ops. He has stared down the promise of AI in property management and lived through the disappointment so far.

Yet he remains optimistic because he understands what is happening.  He helped me understand it in 3 main lessons:

AI ≠ Automation (And I Finally Know Why)

I could never explain the real difference between automation and AI until Kyle put it plainly:

“Once you get a reply, you are now outside the automation.”

That line reshaped how I view every so-called “smart” workflow.

Automation excels at outbound, rule-bound tasks.
It sends the lease-renewal email sixty days before expiry.
It pings a delinquent tenant on day three, seven, and fourteen.
It launches a welcome checklist the moment an owner signs.

These jobs run on if-this-then-that logic.

Everything works well until a real person asks a question you didn’t script.

That’s when you hit what I now call the Reply Cliff.
Automation can walk right to the edge, but it cannot take a single step farther.
The flow freezes. Tasks pile up.

And you’re back in the inbox, babysitting the process you thought would run itself.

That is what excites Kyle about AI.  Not that it’s flashy, but because it finally handles what breaks everything else: inbound.

“Every property manager seems to have the same, you know, 50 questions over and over and over again.”

Now imagine an agent that actually reads the inbox, pulls from a real knowledge base, and answers those repeat questions- clearly, correctly, and without bouncing them to your already overloaded team.

Kyle says when he shares this idea with other operators- an AI that handles the endless back and forth with tenants- they light up.

Because deep down, we all know the real problem isn’t sending messages.
The real problem is surviving the replies.

How Sucky AI Gets Made (and Why Our Favorite SaaS Is Making It)

Kyle tried to patch the Reply Cliff himself. He used a Jotform AI Agent to guide residents through maintenance requests.

The idea was to collect clean data through a natural conversation, then populate the form behind the scenes.

It sounded good in theory. In practice, it broke down fast.

“When people call in, especially if they’re panicking, they don’t start with clean data. They start with emotion.”

The agent should have helped. Instead, it forced everything back into rigid form logic.
And that made the experience worse, not better.

“You either have an agent or a form. Combining the two reduces the usefulness of both.”

If a tool is built to route, not reason, layering AI on top won’t elevate it.
It just reveals the limits that were always there.

This explains why so many “smart” SaaS features feel disappointing.
It is also why some of the biggest AI-featured acquisitions in the industry have underwhelmed once the marketing faded and the reality set in.

The software wasn’t designed to interpret emotion, tease out nuance, or adapt to messy real-world dialogue. It was designed to follow steps.

And when you bolt a contextual agent onto a step-based system, both start to break.

This is why so much property management AI feels… meh.

The Real Fix: AI-First, Outcome-First

Kyle helped me understand that most tech companies claiming to be “AI” today fall into two camps:

• Those layering AI on top of legacy workflows
  
• And those starting with outcomes, then letting the agent figure out how to get there
  

The first treats AI like a prettier interface.
The second treats it like a teammate.

“If you try to use it in a different way, expect it to feel clunky and broken.”

That gave me the mental scaffolding I needed to evaluate what’s real.
But Kyle’s take on value is what will help you navigate forward.

The Non-Obvious Value Kyle Sees AI Creating

When most people talk about the benefits of AI, they mention speed, cost, or even accuracy.

Kyle talked about something different.

He talked about stamina, orchestration, and what teams unlock when the grind is gone.

1. The Patience and Pace of AI Agents

Most people think AI’s superpower is speed.
Kyle thinks it’s stamina.

“The thing that AI has that you and I do not have is unlimited patience… It doesn’t get upset or irritated, or impatient. It will have an hour-long conversation if you want it to, and it will just keep going as if it’s only been five minutes.”

That kind of endurance lifts the floor.
Midnight questions don’t clog tomorrow’s task board. They get handled before you even log in.

But Kyle also explained that it’s not just about workload.

It’s about structure.

“You put guardrails around the AI and let it figure out how to accomplish the goals.”

That clicked for me.

The best people I’ve ever worked with just needed the goal. They could be trusted to figure out the rest.

Now AI can be designed to work the same way.

2. A Team That Loves What They Do

Kyle has seen one PM run 500 doors with 30 people.
He has seen another do it with 12.

The difference wasn’t just technology. It was how they used it.

The best teams aren’t just reducing work. They’re redirecting it.

Toward:

• Strategic owner reviews
  
• Surfacing ops insights
  
• Building relationships that retain clients

They aren’t afraid of AI taking their jobs.

They are relieved it’s giving them better ones.

3. The Big Opportunity (and Threat) to Property Managers

Kyle pointed out that property management is perfectly suited for agentic AI disruption.  Not because it’s broken, but because it is rich with the exact conditions AI agents thrive on:

• High volumes of repetitive communication
  
• Long-term relationships that compound over time
  
• Thin margins that reward every efficiency

“We are inherently a more profitable industry… but also more attractive to disruption.”

And here’s the big insight:
The longer the relationship, the more valuable AI becomes.

Every renewal, repair, and reply leaves behind a trail of context, and that fuels an agent that can eventually anticipate issues before anyone files a ticket.

It flips the narrative.

Yes, property management is a relationship business.
Kyle made me see that it’s the relationship side that stands to gain the most from AI.

The best way to build a relationship business… is to build an AI-first company.

What Happens Next

Kyle isn’t just theorizing. He’s putting this into practice across RPM’s network right now.

That includes a new role most teams haven’t defined yet: the AI Manager.
Someone who tunes prompts, reviews transcripts, manages the agent’s performance, and treats the bot like a teammate that improves with every day on the job.

One of our Vendoroo clients has already stepped into that role.

He calls himself the AI Tzar (and it’s working).

I’d love to find more of you.

If you’ve got that person on your team (or you are that person), reply.
We’re assembling a group. I think you’ll learn a ton from each other.

And if you’re ready to take the first step toward building an AI-first company, come see what the only AI-native maintenance solution can do.

Schedule a demo with Vendoroo.

The conversation with Kyle left me more excited than ever.
Let’s build the future of property management together.

Pablo Gonzalez,

Chief Evangelist at Vendoroo

‍