Do you get the feeling that the hype around AI is quieting down?
At the start of the year, everything was going to be AI and workers would be obsolete. Now, halfway through the year, it doesn’t feel like that at all. It feels to me like the boardroom and exec pressure to “do AI” has hit all the obstacles the experts (and I) flagged long ago.
I’m going to break down these obstacles into two parts. One, the tech itself, and two, what AI is doing, and what it isn't.
The problem is your PMS wasn’t built for AI. It wasn’t even built for front-end operations. Then you've got a handful of point solutions you adopted to compensate for what your PMS couldn’t do, cellotaped into your flow doing different things. And now you’re whacking on an AI tool to “automate it”. And here is the issue: what is ‘it’ that you’re automating? There is no consistency or standardisation, and the workflows that exist are pretty manual and often undefined. That is what is holding back your AI's performance. Bad workflows and bad data.
The legacy systems talk about all the data they have, and yeah, sure, they have lots of data, but you can’t use it. Go into your PMS today, and you will see missing fields, misformatted information, and just generally old data. Your teams don’t trust it, so your AI shouldn’t either.
The other issue is that its master data management (MDM) is property-first, not resident. Makes sense because legacy PMSs were built for accounting and then repurposed to do operations, but effectively, what this means is the most important record is the property, and the residents live within that property. In modern operating systems and CRMs, the MDM is the resident. An important definition if you want to do resident-facing AI.
So that’s your starting point.
You then add on the handful of point solutions, and now you have the leaking buckets problem.
Imagine each system is a bucket of data, and as you pass data from one system to another, it leaks; the more systems you have, the more leakage. The leakage comes from where fields don’t reconcile (don’t match) so get left behind. So, by the time the AI has jumped from one place to another and back again, you’re left with half as much data as you started with.
This is a structural problem. The tech you’re using wasn’t built to do what you’re asking it to do.
Antony Slumbers, Global Keynote Speaker on AI, told the audience at the ARL AI event that you can’t just bolt AI onto legacy systems and expect magic. You need to fix the chassis before you buy the engine.
The engine is AI: the models, the copilots, the automated agents. The chassis is everything underneath—the operational data, the workflows, the processes that generate the information AI depends on.
AI in real estate today is doing the easy work. It’s reading documents, handling enquiries, responding to questions. It’s important, but it isn’t going to transform your business. AI has started here because of Part 1: the underlying technology. Anything truly transformational is limited by how much information AI has at its disposal and how much you actually trust it to do the job end-end forever more (fully agentic).
Again, to lean on an actual expert, Antony Slumbers describes this sort of task as the usual stuff done faster and cheaper. It’s important, but everyone is doing it, so it’s not a competitive advantage.
It’s low risk, but also low reward.
The real ROI comes from changing your operating model. Drastically increasing the units-per-employee KPI. Bringing leasing in-house. Automating maintenance. Things like that.
So, how do you get there? You fix the chassis. The plumbing. The foundation. Whatever you want to call it, you do it first.
This means documenting your workflows and objectives in detail. This then becomes the guardrails for your AI to follow. You standardise your processes across every asset to standardise the data. You aggregate this data to create a real-time view of operations. You then need to align your leasing and property management teams, asset and portfolio managers, and your owners in one connected data layer. You need to make collaboration seamless between them.
Once this has been done, you can start to do what couldn’t be done before. You can scale without adding headcount. You can cut your biggest line items in half. AI has a golden thread of data to build predictive analysis and revenue management engines to deliver longer stays and shorter voids. It prioritises your day so you’re working where the asset strategy needs you.
That’s the stuff that’s going to transform the performance of your portfolio. It’s hard work to build yourselves, but fortunately, I know a rental OS that’s done the heavy lifting for you.
