Why Most PM AI Rollouts Fail – and the 5 Fixes that Work

Derrik Daun:

Alright. Hey, everyone! Welcome to the August VPM webinar. People are starting to trickle in in the waiting room, so… My name is Derek. I am the… I'll be your host today. I am the marketing strategist for VPM Solutions. Pete is unavailable today, so I will be filling in, but our headliner today is Lindsay Liu from Super. And I'll give her the chance to introduce herself here, but if you'd like to, please let us know in the chat. where you are joining from. The chat is going to be open throughout. We also have the Q&A as well. That box is open, so… But Lindsay, welcome, we're very excited to have you.

Lindsay Liu:

Same here. I also see some familiar names, so hey, folks! That burns.

Derrik Daun:

Alright, Bismarck. I've never been.

Lindsay Liu:

Never been, either.

Derrik Daun:

Yes.

Lindsay Liu:

What's it like mid-August in Bismarck?

Derrik Daun:

Hopefully it's not like… I'm in Nashville, and it is… Okay, Bill, same. 90s, humid, stormy, yep, alright, very much similar to here. Yes, the real field temperatures are, like, 105, 110 right now, so… The humidity gets you. There you go, no trees, so no shade.

Lindsay Liu:

No shame, right.

Derrik Daun:

Alright, someone from Mexico, awesome. So, we have people trickling in. We are at 11.32, and we have about, you know, until noon Central Time. So, again, we have Lindsay from Super here to discuss, why most PM AI rollouts fail. We are very excited. To have her, she'll give us some tips and tricks and best practices, so without any further ado, Lindsay, I'll pass it off to you, and thanks everyone for joining. Again, if you do have a question, please feel free to put it in the chat or in the Q&A, and I'll be monitoring that throughout the session. So, Lindsay, take it away.

Lindsay Liu:

Thanks, Derrik. Alright, I'll share my screen here. And, yeah, my name is Lindsay Liu. I'm one of the co-founders of Super. And just quickly on us, we build AI voice, for specifically property management. So, we have seen a lot of deployments of AI. We've seen people with existing tools. We've obviously helped with, hundreds of deployments on our, side of things, and, switches, horror stories, all of that. And so, here to actually just talk about what works and, and what doesn't. From the front lines of it. As far as, kind of, the types of things that we deploy, we're very focused on front office. So, think of answering phones, handling escalations, any kind of tenant and owner communications as well. And so, over the last 18 months, we've had the chance to really work through a number of these with customers. So I'll give you kind of the perspective that we've seen. From that. And definitely, I'm one of those people that likes to say that if we've learned something the hard way, I don't necessarily want anyone else to have to go through that, so sharing, very candidly and openly about that, so that hopefully for the next, we can keep getting better. on that. So I'm going to walk through five, kind of, key themes that we've seen, and we can talk about, kind of, the things that, that don't work, and the things that can help with them, and feel free to just, like Derrick said, drop, some. questions in the chat along the way. Got a question on any of these as we go. So, the first blocker that we see is, honestly, it is just that the process isn't written down. Yeah, a lot of times, it's in somebody's head, maybe it's in a few people's heads, right, depending on, kind of, the departments that they're in, they're kind of like, oh, we do things this way. And the one thing that we definitely see is that even if you have, kind of, your 80% process. there are edge cases that usually aren't nailed down. And so, what ends up happening is it becomes pretty subjective. When it comes to AI, If you want it to follow a specific set of rules on how it behaves, you're going to want to define the criteria for it. So if you leave it to a subjective decision, it's going to make a subjective decision you may not agree with that as well. Since I'll give you a real-life example. We had a customer recently onboarded, and they said, why didn't it route that call to me? They asked for me by name. And then we looked at it, and we said, well, because you… you actually said, if the caller asks to speak with you, then take a message. So you actually said that was the rule. oh, but this vendor should have been allowed to reach me, so here's an edge case, right? Where it's like, oh, but this vendor, like, I should be able to talk to them, in that case. And so what we did was we worked with them and kind of said, okay, well, then we need to update the rule, we need to add some additional criteria here. So if they're asking to speak with you, then we need to find out if they're a vendor with the approved list of vendors that's allowed to speak to you, and then we can transfer. For everyone else, we still want to take a message. So, that's the type of back and forth that we often run into when we're first rolling out, is you kind of, right away, are going to run into all of the edge cases that don't have rules around them. So I would just say, be prepared. For that. And the litmus test for that would be if it's a new employee, think about, kind of, onboarding a new team member, right? And if you don't think they would know how to handle the situation, then you can guarantee that will not be able to, either. And so, these are the types of things that make onboarding and AI very similar, actually, to onboarding a team member, where you want to make sure that the ground rules are really clear. And that, that you're also investing in training. So, the fix to that that we've seen work really well is when we're getting started with Volks, is we usually ask for whatever SOPs we have, any information that already exists. And then, we love to kind of run all of that through AI tools, so, I'm a big fan of Notebook LM, which is provided by Google, so if you have a Google Workspace subscription, it's free. I'm pretty sure they, they, make that accessible to folks outside of the Google ecosystem as well. And you can run all of your own sources, so what it will do is it will limit the scope of what it looks at, just to kind of the sources that So you can put website links in, you can upload documents, you could put, PDFs, so any type of, file. You can even put, we have people that will give us call recordings. As well. So here's a hundred of our call center recordings. Go and look at these to understand how they respond and create that. So it'll analyze all of that for you, and you can start to pull that together into AI-ready SOPs, as well as understanding, hey, where did things break down? What are the edge cases? Here. So, the kind of… one thing I would say for this, specifically, would be work through those edge cases, be ready to refine, and be ready to coach and train just like you would a new staff member, as well. The second blocker that we've noticed is a team one. So, not necessarily for the decision makers that are choosing to implement, but sometimes within the front lines of the team, where the reality, and we can't… I don't think we can avoid this conversation, is people are afraid of being replaced right now, and so they're going to become the blockers, because they don't want to adopt it. As a result. And I think the news is really, really scary right now, right? We're seeing news of layoffs constantly. I think it's really hitting folks in the face right now, so I can really understand where this comes from. But… I actually believe that we're not telling the right story when it comes to that. It's, like, definitely headlining to be able to say, hey, I was able to replace all of my employees with AI. We work on this day in and day out. I will tell you that that is not the case. And so the real litmus test here is to look at the things that your team members are doing. And the things that really drive value for the company, the things that people can uniquely do. So, at the end of the day, property management, real estate, it's a relationship business. And so those are things that should not be, replaced, actually. But what are the things that are blocking your human team members? from having the time to spend on those activities. So, if there's things that you're like, you know, this, you know, for instance, calling all of my owners and doing a check-in with them, I would love for that to happen, but… my team doesn't have enough time because they're… they're just doing the blocking and tackling. Those are the things that we're trying to eliminate. And I think that's the reframe, and that's the way to position it, to your team as well, that this is really here to up-level to them. And so, when we're thinking through that, one of the things that can be really helpful is to actually think about, like, in your job description today, what would you love to have taken off of your plate, right? Make this a co-creation exercise. with the team. What are the things that you'd like to grow into in your role, right? These are the things that, I think are sometimes difficult for people to even see when they're just in the grind of it. Day-to-day. So give them the opportunity to kind of dream in that moment, and find the win-win. That's really the fix here, right, is to look at the areas where you can take the front work off of the team, where can you empower them and find more value-led works for them, and then also, find the opportunity to upskill. into more strategic roles, and frankly, a part of that upscaling is going to be how they work with AI tools, as well, to make themselves work faster and more efficient. As well. And really just bring people along in the training. We've really found that when we have multiple team members involved in an onboarding, that it makes it much more effective, because now everybody understands how it works, they're involved in the training, there's no, misunderstandings around how things should work, or misconceptions around the role of it. It can also be really helpful to establish some internal AI experts, right? So, here are the people that understand the purpose of this system, how it works, they're the ones that are responsible, for it, so that there are people to go to within there. And definitely, I think this is a moment for all of us to really look at How do we redefine roles and job descriptions? What is the… what is the way that a modern property management company powered by AI will leverage its people in its team? Alright, the third one we see is just that you're starting with a trust gap. This is a hard one. At the end of the day, if your owners and your residents already don't believe that you're doing what's best for them, and that you're gonna respond to them, or that, you know, they're having difficulties already, implementing AI into that process is probably going to make them feel like it's just sitting in between. this is a real, Better Business review, from a company. I just found one, you know, it's not that hard to find, frankly, and they were like, I just called, and I called, and I called, and I left messages, and I couldn't get anyone to answer. By that time. you are not going to be able to automate building trust into this process, right? So again, I think this goes back to number two, right, as well, which is where are people going to add value in preventing these types of negative reviews? From happening. And I think that, the key here is that we need to make sure that AI is not seen as being the middleman in between, but it's, that it is actually part of a solution that you're rolling out to make it a better experience across the business. And so, you need to really… it's a reframe, right, for your core customers to make sure that they feel like this isn't just, ugh, you're even making it even harder for it to happen, or you're even removing more of the human touch from the experience. And so I would say one of the litmus tests here before implementing anything would be to, first, just have an honest conversation with yourself and your team that, do you feel like your customers will trust that that you'll get a response to them, that you're gonna do the work well, for them. We've run into this before, where sometimes the issues that you run into are people actually phone bomb, because they can't get… they're so used to not being able to get somebody on the phone. That if they reach your voicemail, they call again, and again, and again, and again. And so they think that doing it more is going to get them more attention. That, to me, is a real clear signal that that trust does not exist. And so the fix here would be, we would say, definitely lead with transparency, and so one of the things there is that you can actually, you know, we work with our customers, oftentimes they'll send out a notice, to both their owners and their tenants, and say, hey, we're rolling out a new system. This is why we're doing it, right? Let them understand what the purpose of it is, and then buy action, right? The idea behind rolling out, any AI implementation is to make your team more efficient and better at what they do, and so this is your moment to be able to improve that as well, that this is part of an improvement that we are making overall. And so, that's, I think, one example of what's worked really well for us, right, with some of these deployments, is you can actually, inform, hey, we're rolling out a new AI voice system, we're doing that so that you can have coverage 24-7, it's going to make it easier for you to get a response, every single thing is involved, and so, you know, really understanding why. And what we've also found is that, this also, you know, you have different audiences. I think property management is one of those really unique businesses where you're dealing with tenants, donors, vendors, they all have different things they're looking for, from you as well. And so, you can also think about tailoring the messages to them. So, for instance, for the owners, one of the things that, can be helpful is to say, hey, we're implementing this. This makes sure that we're answering 100% of the calls that we're getting about renting your apartment, 24-7. And so this is going to lead to better outcomes for you as well. And it makes sure that we have coverage, right, for emergencies. If you need to reach us, here is how, right? So here's the, for instance, we have one customer where they've implemented, essentially, they, they notified all of their owners, if you just say this thing, it will immediately shortcut to me. Right? It will route the call to me right away. So, there's… there's a few ways that I think, you know, just more communication, more transparency, can really help. with this. Alright, the fourth blocker that we've seen is not having clear KPIs, right? Not having clear expectations for what we're looking to get out of it. I think… I think at this point, we're all getting much more, I think because there's so many of these tools, there's less shiny object syndrome, so it's definitely… I've seen a big progression here from a year ago. Even a year ago, it was just, oh, I'm being told I need AI, so I need AI. And it turns out, well, maybe you don't actually have the exact problem that you're looking to solve, or this isn't the priority problem. For you to solve. And so, having really clear KPIs, establishing those benchmarks is great, and then I think also making sure that that is rooted in actual data and in the goals. As well, so the… we definitely don't want to see, like, and even, actually, if we have a prospect that talks to us and just says, well, I know I need AI, I've trained my sales team to say, you gotta go a couple layers deeper and understand, because we want to solve a real problem for you here, and if it's not the right problem, then you're not going to see the results, it's not going to be the right place, right? You spend time. Or, the flip side of that sometimes is we see expectations that might be a little bit of outside of what realistically can happen today, right? Well, I don't want to have any phone calls. I want it to resolve almost all of the inquiries on its own. And so sometimes we kind of sit in between that, and we'll try to workshop, right? Is that reasonable? Is that even the right goal? I think that's actually a really important, thing to work through, is what is the right KPI? Here. And so, litmus test for making sure that you're looking at this the right way would be, you know, what is the end outcome? That you're looking to accomplish, and I really love the first principles style of thinking, which is just take it to the most basic thing possible. Maybe it's not even 80% containment of calls, but maybe it's more… actually, the real thing we're trying to do is to make sure that we get all of our new owner leads right away. And the thing getting in the way of that is that my team, my sales team, for instance, and we hear this is a real thing, tenants will, choose, you know, press number 4 for sales, on the phone tree because they're like, I know I'm gonna get a real person to answer. I don't know if that relates to anyone there, but the sales team is then like, I'm just dealing with, like, leasing and tenant inquiries because they think that I'm going to be the fastest path. So, perhaps, actually, the goal is to make sure that we're routing things the right way, so that the sales team is actually spending time on the phone with prospective, owners. So I think those are the ways that we think through goals, is really, again, first principles, make sure that there's a clear KPI for that. And then, as far as fixing that, right, just get really clear on where the value is in your business, and the things that, that may be getting in the way of, you maximizing those goals. So, focus on making those KPIs the ones that actually are, not just vanity metrics, but the ones that actually influence and affect the business. Fair. Alright, last one, and then I'll turn it over, back to you, Derrik, and, Purple. So, the last thing is really, we see people treat AI as this, as something that doesn't evolve. It's either, you know, it's done, or I train it once, and it's done forever. And the reality of that is that AI is constantly changing, and and we're seeing this, I mean, just this pace of change here, the models, the underlying tech, the way that we're even as technologists building, I know that there are a lot of folks that are vibecoding, as well, you're experiencing just how quickly this change, is happening. If something wasn't working well 3 months ago, it actually may have you know, there may be core changes, to that. So I think there is just a lot of growth mindset, to apply, to this. So instead of thinking it as a, well, this thing didn't work 6 months ago, so therefore I'll never be able to solve this problem, with it, or, I set it up this one way a year ago, and I'm expecting that to continue to work the same way, with different models and adaptations. Later on. That is just, you know, something that I think we're all gonna have to come to grips with, which is that it is a really rapidly changing environment, and make sure that we're adapting our mindsets, to it as well. The one thing that we really, you know, see that that holds people back is that they kind of They spend the time up front, and then they never look at it again. That's great if it's working, but there's probably things that can continue to improve as well. So, you know, try to… there's a saying, right, like, get as much juice out of the squeeze, as possible with the tools that you have on. The other thing that we see that's kind of related to this is that a lot of the platforms out there make you do the setting, part of it, so that's something that, we've definitely seen people, kind of fall short on kind of just getting fully there with their deployments on, so it's super, one of the things that we really do is we have an onboarding team that's, like, very hands-on, they are AI-pumped. experts, and they'll be able to help with reviewing all of those things, and so I would also say evaluate for, kind of, how much of it are you having to do on your own, and again, if you don't have that internal expert, right, that internal AI expert to handle and evolve all of that, then that might be something to evaluate. As well. And so, the litmus test here for you would be, when's the last time I re-evaluated, whether this tool could solve a problem for me? I would definitely encourage, if this problem continues to be a priority on your list, in 3 months, 6 months from now, there have been, I'm sure, massive changes in the way that people are able to solve those problems now, so it's worth continually re-evaluating as well. And then if I have an existing, tool in place, when was the last time I reviewed and adapted and actually thought about, hey, is there a way I can actually get that from 80% to 90%, right? So there are definitely opportunities to continue to evolve that. It's not just the, oh, it's not working, or it didn't exist well, so it doesn't, it can't be done. And so, the fix here really is, treat it as iterative. It's constantly evolving. For the existing tools we have, keep training your data. We have this awesome one of our customers, ran all of his call transcripts through, through Claude, and he came back and said, hey, I noticed that these are the five gaps, I've built knowledge bases for them based on those gaps, and, I'd also like to add X, Y, and Z. As well. What an easy way to actually just have AI do the hard work for you on looking through where the training needs to be updated. So these tools are all at our disposal now to be able to really understand, how things, can improve. And then also the AI tools are great for research for, new tooling, that's out there as well. They're constantly following the internet. And I think they just, you know, they're more on top of it than maybe another person. Could be. And at the core of all of this, I think, you know, just to quickly summarize. when I look back at all of these, kind of, blockers, and Fixes, they're not really… Right. It's about things that this business has always been about. It's process, it's people, it's managing your data. And I think these are things that property management are actually, as an industry, is particularly in a good position, to capitalize on, because these are the things that make for a good property management, company. So, I think it's a really, really exciting time, and would not want people to be discouraged just because, you know, a couple of experiments didn't work out for them, but I would say, you know, get to the core of it, look at, the things that you're looking to improve, and try to find ways to, To iterate, on those as well, because they're usually things that, that if you kind of get to that first principles, approach, you can find a solution. Or… Alright, I will stop. There, turn it back over to you.

Derrik Daun:

Perfect. Thank you, Lindsay. A lot of good information, there. So we have about 5 minutes left. If you do have a question, please put it in the chat. I think one thing that really stood out to me, and I think it's a common perception, is this set it and forget it. Like, hey, we have it installed, we have it using. And now it's just gonna work without us needing to be hands-on. And not iterating. And that is definitely something that It's for efficiency, but it's always evolving, as Lindsay said. I do have a question, though. Because you have these 5 blockers, and I think as we were going through it. I'm sure most attendees will say, oh, I know that blocker, I feel that blocker, and they're not all mutually exclusive, you can have multiple blockers at one time. Are there one or two of the more common blockers that you find that are the most resistant, or the most challenging for people to get started?

Lindsay Liu:

It really is, I think, the first one on having the processes really clearly laid out, right? I think, you know, if you don't… the AI tools right now, the reasoning capabilities are pretty good, but It's not in your head, it's not reasoning the way you would reason, necessarily. And so, if you leave a gray area, it is going to fill that gray area with something. So that's just something that we know about how these tools work, is it always wants to give you an answer, and there's… there's actually some technical reasons for that with the way that the tools are benchmarked, where, actually the benchmarking kind of, reinforces the, let me just guess versus saying I don't know, because, hey, if I say I don't know, I have a 0% chance of benchmarking against that, but if I throw an answer against the wall, maybe 1 out of 100 times I'll get it, right? So that's actually why, a lot of the tools are… are… they over-index, into that. So, the processes are definitely the… the biggest sticking point. They're the things that usually are where we see holdups in onboarding, and it takes a little bit longer to work through all of those. And then again, it's not just a one-and-done situation, it's not, I gave you all the processes. Here you go, it's now we've seen there's these, these scenarios that, we didn't have a solution for. In the process, how should we handle that? So, making sure that you're having an iterative conversation about it.

Derrik Daun:

Absolutely. I think also, too, is, like you said, it's… you know, there is that fear of, I don't know it, but also, you know, the more you're in it, the more you're gonna learn. And, I think that's important for people to know as well. If you have a question, please put it in the chat. We have just a couple minutes left. So I'm going to, just share my screen and give everyone a chance to see what we have coming up here. So if you'd like, our next webinar is next month. Tuesday, September 15th, we will be with Robert Block from Rent Recovery, so you can scan that QR code as well. But again, if you have any questions, please, please do so. We do have a couple more minutes, Lindsay and I will stay on. But this has been a very helpful, and I think, you know, people continue to hear AI And then sometimes you get so overwhelmed that you just Decide to block it out yourself, but it is important to… to at least take those steps for you and your business. Okay. All right, well, no other questions, but any final thoughts for you, Lindsay?

Lindsay Liu:

No, I think I'm just, excited about seeing just the volume of people that are willing to, to… give some sort of experiment a go right now, and kind of, you know, whether that's, you know, people that work directly with us, or just in the industry at large, and really seeing so much, activity, and so I think we're gonna see, we're gonna continue to learn, and I think it's important that, as a community, we continue to share what's working and what's not working. It's the only way to improve and get better.

Derrik Daun:

Absolutely. I know you had, HireSuper.com, I believe was your company website, so… Yeah, I mean, if anyone wants to reach out to you or the team and just have that conversation, highly recommend that. So… Perfect. All right, well, we are at time. Thanks again, Lindsay. Great stuff. Hope everyone enjoyed it. And have a great rest of your day, and we will… we'll see you next month.

Lindsay Liu:

Thanks all. Bye.

Derrik Daun:

Thank you.