Implementing AI is an iterative process, from leveraging current tech to continually assessing evolution in the space, and taking advantage of more powerful tools as they become available.  I had the pleasure of talking with Cat Allday,  Vice President of Product Management at Appfolio, on how they view technology adoption, product innovation, and evolving their own offerings over time.  

Brent Williams: Hey everybody, my name is Brent Williams. I am Chief Insider at Multifamily Insiders, and today I have the pleasure of talking with Cat Allday, who is Vice President of Product Management at Appfolio. I'm looking forward to talking today about the technology revolutions that happen seemingly more and more often.

Every time there's some new technology revolution, there are a thousand different ways that companies can take that and then improve some process, flow, functionality, or people's lives in some way. And so I want to hear from Cat about how her and her team then use these new technologies to then enhance property management operations or improve the work processes of property management professionals.

Cat, thank you so much for joining me today.

Cat Allday: Thanks for having me. I'm really looking forward to this conversation.

Brent Williams: Right now, of course people are talking about AI first and foremost, but I'd like to actually back things up a bit because Appfolio has been around for a bit, and so it's been through several of these technology revolutions. I think that we can learn from how these earlier technology rollouts set the stage for what you're doing now.

So let's rewind the clock a bit and talk about what you've done in the past when big changes have happened on the technology front.

Cat Allday: Yeah, Appfolio was founded around 2007 and at that time a lot of property management software solutions were not hosted in the cloud. And so Appfolio started building a software product that was a hundred percent hosted.

And so we really took advantage of what was going on in technology at that time, which was the cloud. Fast forward a few years. The next sort of technology revolution was mobile. We all started carrying these things in our pockets that had access to the internet. And not only did in our personal lives, but at work we also wanted to use them.

And you think about property management, these people were on the go all the time, whether they're on site doing a maintenance call, doing a tour, and having their software available to them on their mobile device was game changing. And then, starting in sort of the 2020 timeframe, AI was really something that was on the horizon, and we wanted to pay attention to that technology revolution that was coming.

It was impending, and I think we already had a track record of paying attention to what was going on in the technology space and how we could solve problems for property managers with those new technologies that we were going to be enabled with. And AI was no different.

Brent Williams: How do you marry the legacy systems and having the cultural mindset to then adopt new technologies? Because I think a lot of companies are built a certain way and now they have these new things coming on board. How do you get everybody on board with that and not have people scared out of their minds with these new opportunities?

Cat Allday: Yeah, I think there's a couple of ways. You can't just build technology for technology's sake. You actually have to solve a real problem that people have. And we've always been focused on using technology to solve a problem, not to just build the newest, coolest thing. And I think that's been an ethos that has really helped us make sure that the technology that we're building actually provides benefit to our customers. And that's an important thing. And as somebody who's worked in tech for a long time, it's really easy to get excited about new technology. And you could build something that actually doesn't provide any value, but it's really cool and that is definitely not what we want to do.

Brent Williams: You know, it's funny, so we just had a multifamily demo day event which you all were a part of. And one of our panels was talking about building the business case for different technology. And it wasn't about just what is fun, what is cool. It's a, as a matter of how you quantify things, how you establish the actual true tangible value of running these different technologies.

So very timely. Okay, so let's get now into AI. Tell me about your first iterations into AI, how you viewed that from your processes.

Cat Allday: Yeah, so we were exploring building AI to solve certain problems around workforce efficiency. And we actually acquired a company called Dynasty.

And that's how we got our first what I would call digital employee product, which was Lisa, which is a digital leasing agent. And one of the things that we believed was that property managers were already outsourcing some work. They were either outsourcing them to call centers, some of them were outsourcing overseas, and these felt like prime opportunities to use technology to solve that problem for them.

And so that's where we started thinking about what were some jobs that people were already outsourcing and could AI help solve those problems? And so we started with a leasing agent, the digital leasing agent, Lisa. And then smart maintenance was another product that we built around the maintenance workflow.

And so those were those first two forays. We also were doing things around invoice processing with our smart bill entry. Again, thinking about the tedious work, right? So entering bills. It's just routine work. If we could use technology to help people by uploading an invoice into the system, having the technology be able to understand what was on the invoice and create the bill and then the team could then just review the bills and do any categorization or anything that they needed to.

But that data entry, heavy work was taken off their plate. Those were our sort of first forays into a AI.

Brent Williams: A couple of things on that. For the first one I really like that approach because, when you think about the people are getting used to AI now, like it's getting to be more pervasive and people are like, okay, I kind of understand how it's working now some more than others.

Yeah. But the idea of saying, you're already doing these things like they're already be done offsite or they're already outsourcing them in some way. That seems to be a much easier way to get buy-in and adoption for something like that, versus something that's the key core fundamental of what they actually touch their hands on every single day.

And you brought up something interesting because you were talking about how with invoicing, for example, and then the team members can still review certain things. It seems to me that a lot of people have this all or nothing attitude with AI.

Cat Allday: I think it's about augmenting somebody or even supercharging them, right?

So I don't have to do this big list of things. I'm going to focus on these three things that I can really add value into the process. I think that's the way I like to think about how we are helping product property managers be more effective in their jobs because they're able to focus on the most important things versus that tedious, mundane, routine, kind of work that nobody really wants to do.

Brent Williams: Yeah, absolutely. And besides, we've thrown so much, especially at the onsite teams that they can't possibly do everything. And so it's so funny. We keep hearing about how an obscene number of calls are not actually picked up at property still.

For years we talked about this idea of going back to basics as if we can just convince them to answer the phones and that's the really the problem. But the reality is that they're just busy. We have all these new systems that we're throwing at them and they don't have time, so things are going to get left off.

And unfortunately, sometimes there's those calls. So if we can take those things off of their plate, then we're assisting them versus replacing them.

Cat Allday: I love the analogy of when you think about the receptionist of the old days, right? That were set up on a desk, answering the phone nonstop.

Now, if you could take that person and instead of them answering questions and answering the phone and being disrupted in their work, and you can instead say you're going to go into the field and you're going to actually engage with the residents. You're going to build community, you're going to do all these things that only a person can really do right. And we're going to let technology handle some of this other stuff. That to me is what's exciting. I like how this industry can change. You can actually have people building relationships, building community, creating a great experience for your residents versus like just being a receptionist,

Brent Williams: Yeah, exactly. We leverage what makes us human, and then still hand off the things like you're talking about. It's also some element of this is horrible work. Like, I don't want to have to do this data entry or whatever the case may be over and over again.

So you talked about some of the initial steps into AI and let's fast forward a little bit where you stand now.

Cat Allday: Yeah, so in the beginning, we were really focused on traditional AI development, machine learning. And as the technology has advanced and as you are well aware, it's in the news everywhere.

The advancements in AI are moving so fast, it's really hard to keep up. However, about 18 to 24 months ago with the launch of these large language models and generative AI, we were already poised to take advantage of that. Because we've been thinking about AI, we've been building it using these traditional models, and now we had access to this super powerful natural language models that allowed us to really think about how can we take advantage of that and supercharge resident communications the maintenance workflow in a way where people can engage with the AI in a very natural human-like way.

Brent Williams: And just out of curiosity, so you had your own AI systems, and now these big major players came out with theirs? How did those two things play together? Did you just replace, did you augment it? How does that work?

Cat Allday: Yeah, so we want to be agnostic to the large language models out there.

So, OpenAI has something, Anthropic has models, Google. All these big players have these models and so we want to take advantage of the best for the use case. And I think that's where we've built this thing called our AI factory, which is this foundation which allows us to say, hey, for this use case, this model is actually better.

And we can build guardrails around how we use those models that are very specific to the real estate and property management domain because we have that experience. So taking these generic large language models, creating guardrails and rules around domain specific information allows us to really build this sort of very targeted AI solution.

And so that's what we've done. We're leveraging all different kinds of models. We're taking the best of them for the specific use case, and then we're applying domain specific rules of engagement around it, if that makes sense.

Brent Williams: Is creating those guardrails an iterative process, I'd imagine the first guardrails you set up, you start seeing exceptions and you're like, oh wow, we need to tweak things here. How does that work?

Cat Allday: Yeah, so we are always monitoring what the AI is doing, the results it's giving. We also look at how our customers are engaging with it, what types of questions are they asking, what types of tasks are they trying to get it to do?

And by monitoring that, we learn a tremendous amount. And we can then go in and say, oh, for this question, here's the best model to use. Here's the best set of rules that we want to build around it. So if somebody were to ask a generic question about something not real estate related the AI will return saying say, hey, I can't help you with that.

That's not in my purview. So we definitely want to make sure that we're being more specific. If you want to go ask general questions, you can go out and use ChatGPT. But when you're in our product, we want it to make sense to the work you're trying to accomplish.

Brent Williams: Do you use one AI to monitor the other AI to make sure they're not going off the rails?

Cat Allday: That's a good question. I have to ask the team. I'm sure that there's probably some of that analysis that's being done using AI. That's fine.

Brent Williams: Okay so let's talk about some stuff that you're doing now. When we talked before, you had mentioned something about workflows. So can you walk me through that.

Cat Allday: Yeah, so we released a product called Flows, which is about how do you codify your business operations for certain functions. So let's take a delinquency. Everybody has to deal with delinquencies. You probably have a standard operating procedure on how you deal with delinquencies.

When you send a notification  .The frequency of those notifications and then what happens when you need to go to collections or, whatever that flow might be. And so we've built a an application that allows you to codify those specific workflows and then apply AI to that as well to potentially take over some of the work for you.

So you already know that you want to send out automatic reminders. Or you want to move something to a certain person. All of that can be automated through our Flows product. And we're looking at some very like specific common flows that are used across the industry. So we have a delinquency flow, lease renewal, an inbound lead nurturing flow.

And then we're continuing to build out more flows. But, I think what's exciting is in the future, you'll be able to not only assign that work to a person, but also assign some of that, those tasks or activities to AI to help you.

Brent Williams: Oh, so you're talking about like an AI agent at that point?

So anything else that we should be on the lookout for the future?

Cat Allday: I think, we are all always looking to drive efficiency for our customers, but not only driving efficiency, but we actually want to drive performance outcomes for your business.

So not only do we want to take work off people's plates, but we actually want to be driving incremental improvements to your business. And so we're partnering very closely with our customers to understand what's important to them. And I think that's where the power of AI really comes in. It's not just about taking work off people's plates, it's actually about fundamentally changing the outcomes of your business, whether that's by driving down costs or finding new opportunities for you to make more revenue or improve your resident experience, which drives up resident satisfaction, which leads to people staying in your properties longer.

I think that's really the power of AI and that's what we're always focused on, is how do we actually drive those outcomes for your business. Not just improve efficiency.

Brent Williams: Yeah, we can do the things, but now how do we quantify that piece?

Cat Allday: I had some data that I thought I could share just around how the industry is thinking about AI. So there's a new benchmark report coming out around property management and the use of AI. And we are seeing the use of AI increase year over year. So it was I think 21% in 2024.

And in this year, this month's report, it's already at 34%. So 34% of property managers are encouraging their teams to use AI. And the most common is resident communication at 60%. So people are recognizing, hey, it's important for me to make sure I'm responding to my residents in a timely way that drives resident satisfaction, and we're seeing people applying AI to that resident communication.

Brent Williams: Interesting. I'm actually surprised it's not higher because it seems every supplier's going to have some element that are inserting AI in different pieces. So I would be, I'll be surprised to find anybody at this point doesn't have some, maybe they don't realize it's happened.

Maybe it's behind the scenes that something is happening, churning there. We actually have a technology adoption survey that we are working on right now. And so we assess similar things about how technology adoption has progressed in different ways.

So I'll be curious to see how our data matches up with your alls.

Cat Allday: Yeah, I can't wait to hear. Because I think it's a very interesting space where people are trying to decide how much they're going to defer stuff to AI versus how much they want to hold onto. Absolutely.

Brent Williams: Well, Cat, thank you so much for joining me today.

Cat Allday: This has been fantastic. I've really enjoyed it. Thank you for having me, everybody.

Brent Williams: I hope you enjoyed my conversation with Cat, and we'll see you next time. Bye.

Cat Allday: Thank you.