It worked. That was the problem.
"The emails are great and really informative... But what I really want is to speak with my property manager."
That's a tenant at OC Real Estate in Adelaide, replying to an automation that was doing exactly what it was built to do.
OC manages 4,000 properties. A while ago, the team rebuilt how new tenants get settled in. The old way was a pile of information on day one. That's the same day someone is juggling a moving truck, picking up keys and wondering if the power's on. Now new tenants get a series of emails across their first couple of months instead. How the portal works. What to expect at an inspection. How maintenance works. All the things property managers explain over and over, every single day.
It did its job. Repeat questions dropped, property managers got time back, and every tenant got the same information whichever office they rented through. On paper, a clear win. Then that reply came in.
"We'd solved the operational problem," says Teegan Edwards, OC's Systems and Projects Manager, "but we hadn't fully solved the customer problem."
A broken automation gets noticed fast. One that works perfectly and still misses the point can run quietly for a long time.
We judge automation by what it saves us.
Fewer calls. Fewer repeat questions. Fewer hours on admin. Those are real wins, and most agencies measure their automations exactly that way.
Teegan is honest about how the thinking drifted at OC. It started with the right questions, about working more efficiently and improving service. Over time they narrowed into something smaller: can we automate this, can we add it to a dashboard, can we turn on AI email replies?
"Every problem seemed to have a technology solution," she says. "And to be fair, a lot of those solutions worked."
Nobody did anything wrong. The problem is the scoreboard. When an automation is judged by what it saves the agency, nobody is checking how it felt on the other end.
They had the information. They wanted a person.
OC's tenant wasn't short of information. They had more of it than ever, delivered on time in neat, digestible pieces. What they wanted, Teegan says, was reassurance, and to know "that there was a person on the other end listening."
She felt the same thing herself when she bought her first home. Her broker sent the mortgage documents with a short note. "Give me a call when you've got 30 minutes there and I'll walk you through it." Every answer was already in the paperwork, and she could have read it on her own. She called anyway. What she needed wasn't more pages. It was "confidence and reassurance", from someone who could explain it.
Owners are telling OC the same thing. Over the past year, OC asked more than 800 new owners how they'd like the agency to contact them. Just over half chose a phone call. 31% chose email and 18% chose SMS. Email still matters, for serving notices and putting decisions in writing. But the most popular answer, by a distance, was a person on the phone.
"Sometimes we assume that customers want less human interaction because tech exists and it makes it really convenient for them," Teegan says. "But often what we found is what they really want is easier access to the right human at the right time."
You know this from the other side of the counter. Marisa Reeves runs client success and operations at SoCo Realty in Perth. She described trying to reach your bank and getting "about 10 rounds of troubleshooting with AI before you can actually talk to a person." If that frustrates you, it will frustrate an upset tenant even more. Her rule: whatever you automate, there's always a clear way through to a real person.
Not can we. Should we.
At a recent conference, OC's team watched a demo from the US. One AI bot took a tenant's call, asked the questions and diagnosed the issue. Then it handed over to a second bot, which created the work order and engaged the tradie. No property manager. No person at all.
"Technically, it was really impressive," Teegan says. OC's answer was still no: "That's not the roadmap that we've chosen."
That isn't a rejection of technology. OC still automates for consistency across every office. What changed is the question it asks first. "Instead of only asking, can we automate this? We now ask, should we automate this?"
You can't answer "should" from inside the agency. You have to hear from the people on the other end. That's what The Method means by reviewing feedback: every comment "is a window into how your maintenance process really feels from the outside." OC's whole rethink started with one reply from one tenant.
That's the test worth borrowing. Not "did it save us time?" but "did they get what they needed?"
Two exercises for your team.
Anna Pengelly, Tapi's Managing Director, hosted the panel and closed it with a few exercises. Two of them fit here.
Map it. Take an owner from prospect all the way to renewing their management agreement, and write down every step. Then mark where automation helps both you and the owner, and where a person makes the difference.
Read your escalations backwards. Pull every escalation from the last 12 months and ask what triggered it. A template? An email that couldn't carry what you meant? "I think you'll find that a lot of them are about how we communicated the issue," Anna says, "not the issue itself."
Then take the automation you're proudest of and ask two questions. What problem does this solve for us? And what was the customer actually asking for? If the answers match, leave it alone. If they don't, you don't need to switch anything off. You need to add a way to reach a person.
So here's one for your team this week. If a tenant replied to your best automation tomorrow, what would they say?
More on where automation ends and you come in: The 36th reminder: when to stop chasing and pick up the phone
