The AI Trap for Property Management Owners
Sam Eddinger is a former nuclear engineer who personally built hundreds of LeadSimple processes for his property management company. Then he realized those processes had turned him into the bottleneck.
That's the tension I wanted to dig into when Sam came on the podcast. He owns Ironclad Property Management in Connecticut, runs a nearly fully remote team, and thinks about his business the way you'd expect a physics guy to: in second- and third-order effects. We've hiked a canyon together and argued at conferences, so this one got honest fast.
The question underneath the whole conversation: as the owner, should you be the one going deep on AI and vibe coding? Or is that the exact trap you spent years climbing out of?
"Let Go of the Vine" Runs Straight Into AI
Sam pulled a line from Traction that stuck with him: "let go of the vine." Holding on so tight that nobody else gets to be successful in their role.
He's lived the cost of not letting go. He built all his own LeadSimple processes ("I have processes for my processes," he told me) and it worked, until it didn't. He became the single point of failure. When Ironclad moved to GoHighLevel for their CRM, Sam made a deliberate choice: he would not learn it. At all.
I've made the same call. When we rolled out Property Meld, I decided I wasn't going to learn it, and I still can't do much more than log in. Six years earlier I'd have been all over it.
Here's where it gets hard. AI isn't Property Meld. Sam put it plainly: going back into the day-to-day to build and maintain systems is "in conflict with my growth as an entrepreneur and leader." He's probably better equipped to vibe code than anyone on his team, which is exactly the problem. If he picks it up, he suspects he'll never put it down.
So I asked him two questions I'd ask myself.
Do you believe AI or vibe coding could solve your current constraint at Ironclad? His answer was no.
Would you just enjoy messing around with it? Yes, but he's afraid it would suck him back in, away from the leader he's trying to become.
Where Sam and I Landed Differently on AI
Sam's logic is clean: being the AI guy won't help him work less, and working less while making more is the whole point. Hard to argue.
Here's where I pushed back a little. I don't think AI is like learning how your office printers are networked. It's a once-in-a-generation shift (closer to the internet or social media) where being genuinely fluent can throw off an outsized return for an owner. Sam wants to be fluent for his kids and for the joy of contributing. I'd add a business case: this is one of the rare things worth staying close to even after you've delegated everything else.
But I'm not telling him to go bury himself in Replit either. If you get your workweek down to ten hours and you want to spend another ten or twenty playing with AI because you love it, that's the reward for building the business in the first place. The trap isn't touching the tools. The trap is building something integral to the company and waking up as its full-time bug-fixer.
One place AI clearly earns its keep: proving our value to owners. I've wanted this since day one: a way to surface everything we do in the background. Rent collected, showings completed, maintenance resolved, code violations handled, evictions managed. Owners see maybe 5% of it. Now you can wire into the Meld, Buildium, and LeadSimple APIs, run a little logic, and notify the owner every time you do something good for them. That used to be a fantasy. Now it's a morning in Claude Code.
The Single-Person Departmental Pod
Ironclad is organized in a way Sam calls "single-person departmental." Each pod runs 300 to 400 units with five people: a client success manager (client-facing), a maintenance coordinator (vendor-facing), a resident experience specialist (resident-facing), a field inspector (the eyes on the ground), and a leasing specialist who also handles evictions and Section 8.
The design is deliberately a little adversarial. The maintenance coordinator advocates for vendors, the client success manager for the owner, the resident specialist for the tenant. That built-in tension, Sam argues, produces better outcomes because each person is genuinely fighting for their constituency.
The owner always talks to their client success manager, by name, with a direct phone number. Other roles can email the owner, but the CSM is CC'd on everything so nothing happens behind their back.
Sam was straight about the downside, and it's a big one. The client success manager has to control everything happening for that owner and lead the four or five people in the pod: one-on-ones, standups, vacancy meetings. That's a punishing combination, and it demands a rare person who's strong at both operations and leadership. In a straight departmental structure, you'd have a dedicated manager providing that leadership layer.
Why I Went the Other Way: Trust Over Structure
We organize differently at RL, and the reason comes down to one word: trust.
We have a single-person pod (the property manager, who's basically Sam's CSM) and everyone else sits in departments like leasing and turns. All ~380–390 owners are assigned to one of three property managers who own communications, renewals, and escalations. Mostly, they exist to build a relationship.
When an owner loses trust in a property management company, the numbers stop mattering. They just leave, because the company feels unpredictable. You can't fully trust a company, but you can trust a person. You even learn what they're bad at, and that predictability is its own kind of trust. What kills it is six people touching your property when you don't know half their names.
Sam pushed back here, fairly. Because RL has in-house maintenance, our property managers can actually own decisions. Ironclad outsources maintenance, so their CSM needs the internal tension with a maintenance coordinator to justify decisions to owners. His concern about my model: a property manager who's gathering information but isn't part of the decision might struggle to defend it. There's no perfect structure, just tradeoffs you pick on purpose.
I'll give the structure some credit, though. As of April 6, 2026, our year-to-date churn has never been lower: unit churn tracking toward 10% for the year, easily our best ever. Some of that, I think, is owners now calling their property manager instead of their property management company. Small difference. Big goodwill.
The Idea I'm Stealing: Monthly Performance Reviews
The thing I most admired (and I told Sam this is a genuine weak spot for me) is his performance management.
He walked it up the ladder: annual reviews, then biannual, then quarterly. Every cadence was too slow. People have good days and bad days, and you want to know much faster whether someone is actually succeeding. So Ironclad went monthly.
The mechanic is dead simple. A two-minute email per person: I rate you a 1 (not meeting expectations), 2 (meeting), or 3 (exceeding), and here's why. It's proceduralized, tracked in a spreadsheet, and two consecutive 1s triggers an automatic PIP. Nobody's surprised when things aren't working, because feedback shows up every 30 days.
I was honest that this isn't me. My instinct with a team member is closer to: you're either doing great, or I quietly want to fire you, and I don't want to manage the in-between. That's the engineer in me in a bad way: wanting the machine that produces results without the messy human middle. Sam's system is the discipline I keep avoiding.
The Real Takeaway
The line from this conversation I keep coming back to is the one about how hard this business actually is. Here's a nuclear engineer with hundreds of processes, a bought-in remote team, and real leadership chops, and he hasn't yet crossed 400 doors. I told him, and I meant it as a compliment: if it's this hard for him, then managing other people's rental properties is genuinely, structurally hard.
Which is the whole argument for being intentional about where you spend your own attention as an owner. The AI question, the org chart, the review cadence: they're all really the same question. What is your highest and best use, and what are you doing out of habit because you're good at it?