Inside the Funnel with Mallory Sheibley — Designing AI for a people-first multifamily industry

(upbeat music) Welcome back to Multi Family Unpacked. I'm Hannah O'Leary, producer of the show. Before today's inside the funnel conversation, we have one important point. AI alone can save time, but real impact comes when it's paired with the right operating model. And that's where centralization matters. In this next clip, Tyler Christiansen, CEO of Funnel, breaks down why AI and centralization need to move together. And why, one without the other, won't deliver the outcomes operators are actually looking for. Let's check it out. (upbeat music) - It's a really bold statement, so we said you can get cost savings, but one of the important learnings we have is that AI alone will not deliver cost savings for you. If you're not centralizing, all AI is gonna get you as time savings, right? You're not gonna actually save any money. And maybe that's worth it. Maybe you should do it anyway, probably should, but don't expect cost savings. - Centralization must be including automating the ever-dendit tasks that are so perfect for the new operating model. If you centralize them and you automate them, it will be a better consumer experience, better agent experience. And ultimately, there will be genuine cost savings there, but you have to do both of those elements. Cool, and they're going to realize that they like the technology, but they didn't get the outcomes they were hoping for. Yeah, you wanna keep the technology, but what you're missing was the operating model. You're missing aggregating the work, right? Like a two-person accounting firm is not gonna be able to fire anybody because of AI. You know, but the big four are, right? Like those ones are gonna be able to do it. (upbeat music) That idea that AI works best when it's designed to support a centralized operating model and empower people instead of replacing them is exactly what we're unpacking next. In this edition of Inside the Funnel, I sit down with Mallory Shibley, lead product manager of AI at Funnel. She's one of the key team members shaping AI and automation throughout our product suite. She shares how we can think about guardrails, handoffs between humans and AI, and what it really takes to build technology that works with teams in a people-first industry. Let's get into it. (upbeat music) Mallory, thank you for joining us for Inside the Funnel. We're so glad you're here. If you were explaining what you do, which is incredibly technical, how would you explain it to a five-year-old in your family? Yeah. Maybe a middle schooler. Let's make them a little older. Fair enough, I think it's a good call. I also love this question because I think it's good for a middle schooler, but this is also the way I'm now gonna explain my job to my friends who have no idea what I do. I'm a product manager and specifically at Funnel, I'm building AI and automation products, which adds another layer to it. So if I was talking to a kid about it, I'd probably start relating apps. You use a phone all the time, you have apps that connect to different pieces of information and the way you interact with them. Well, there's an engineer that you might imagine on the backside of that who's coding away and writing those things and telling the computer what to do. I'm one step even behind that where I am helping the engineer decide what it is that they're going to do, understanding what the person using the app wants to do next, something that they're a problem that they're running into. All gather all that information and then kind of conduct the engineers who are working together on that and kind of connect together whole teams and hopefully deliver to you an app that you like to use a game that you might like to play, making Candy Crush as least buggy as possible. - I love that. It's a good translation layer between end users, what they need, good AI to help some fine departments and our engineers who want to be in the weeds in the code. So. - Absolutely. - We're glad you're here. For a period of time, I'm not trying to call you out. You love funnel and we are so glad you're back. Let's just say that up front. But what made you want to boomerang back around? What drew you back to funnel? - Absolutely. And yeah, focusing on the best part of it, which is that I am back here now. It was one of those, even when I was leaving my last couple of weeks, the messages that people sent me about how they really appreciated the relationship that we had while I was working here. It was one of the hardest things, oh my God, I'm making a mistake and I was. And so I stayed in touch with various people that I worked with at funnel so much more deeply than I thought that I was going to and keeping that connection going. And also just, I knew even when the time that I was here, what a special place funnel is, the culture that we have, the way that I work with so many different teams. I mean, as I mentioned in my last answer, I'm working with engineering, I'm working with design, I'm working with the customer, I'm working with everybody. And the relationships that funnel, the way that we set up relationships amongst those groups is unique. And I knew that, but I think I didn't appreciate it as much. And it was just one of those things looking back, you're like, wow, not only is it that I get to really enjoy the people I work with, but it's also the way that, you know, from leadership down, the way that, you know, Tyler has his team's work together, the way that leadership talks to each other, shares and is transparent with us, you don't get that everywhere. And on top of that, the products that we are putting out there, truly getting to see the NPS scores we get back sometimes from people who are online leasing, being like, this is the easiest thing I've ever done to be able to apply to an apartment in four and a half minutes. Sometimes it sounds silly, but then you're like, wow, that's actually something people have to do every single year, sometimes when they're moving. And so you're really making an impact and actually having a good time while doing that, you do not find that anywhere. - Like I said, we're so glad you're back. You've worked in AI and finance, consulting, and now you're in multifamily. What did you learn from those earlier industries about how humans and AI can work together or humans and technology are better together and can complement each other that you carried over into your mindset and how we work here at Funnel? - And the first thing I'll say is also something to think about, I know it feels like it's been such a long time that we've had generative tech and this AI that really can, you can ask it to do things for you and it will do them. When I was working in AI and consulting, working on pretty primitive is how we would describe it now, chatbots. And then in finance, it's especially with how regulated in an industry like that is, there really wasn't a replacement for humans. There was nothing that it was necessarily doing on your behalf. And I was a little bit because it couldn't, but also a little bit because people were like, that would be really unreasonable. I'm an expert at my job. I'm the human, I know best I have all of the context, but even as we've now gotten smarter and as I look back to see what my peers are doing, I think they're also still grounding in that we are, this is a helper tool. Just like really any technology is, AI kind of superpowers that as to what it can do for you, but it is supposed to make my life easier. It's supposed to be an add-on to my life. It is not supposed to be a second me. It could be a second pieces of me. It could be a second small pieces of my workflow, which is awesome. And it is supposed to boost every human and reduce mental load. The best use of automation AI is to get rid of repetitive tasks. You can do what you as a human are best at, which is relationship building and leaning on that emotional side of you. And also just being very smart. That is the best place to marry technology and AI in particular to improve your life and make your life better, but still get to be creative yourself, still get to be hands-on and build yourself where it makes sense just for the boost. - Thinking about crossover between industries that you've worked in, I think about the regulation and finance, I think about fair housing and multi-family and just making sure AI has some guardrails in place to make sure that when it is able to run in a workflow, it's running in a way that keeps everyone safe and where it can really run, it has the freedom to do so and really aids teams on those repetitive tasks. No one wants to copy and paste a pet policy 12 times. Funhole often talks about AI as an assistant and this builds on your last question what you talked about there also. But from your seat and building our roadmap with AI, how do you design AI so that empowers leasing team members rather than sidelineing them? - Yeah, absolutely. And I think this is really the most important question of the year and as we expand out funnel and also just AI in general. We've always had this stance that we want to be an AI assistant. We are empowering our leasing agents and all the agents who work within funnel every day. Leasing in particular, there is such an emotional relationship building aspect of it that is so important and we don't want to take the heart out of that. We are boosting crews and if I think about myself and how I'd want AI to be helping me and how it is helping me in my workflows, it's helping me bounce ideas off of things. It's helping me be creative quicker. It's helping me be technical quicker. But it's not doing the heart of my job because it can't and I think it's the exact same thing for most industries that is clearly the best way to do it. And also one other thing to note on the roadmap thing specifically, our product is B2B2C. So of course we spend a lot of time with our direct management. But we also make a very big point to spend as much time as possible talking to leasing assistants and leasing agents themselves to learn about what their paint pointed directly and to their question about how is that getting back into our roadmap. That's how we are spending a lot of time with them. We're looking at the data, the usage of our app and then all of that is feeding back into it along with the information about the people who are making the buying decisions. So same thing, trying to balance all of those different moving parts together. It's important to think about the end users. People who log in in the morning and then log out a funnel at the end of the day. When you think about, I mean there's a lot of hype with AI right now. So I feel you probably get asked a bunch of questions on weekends or with your friends. But when you think about your role here at funnel and how we balance that excitement of the possibilities of what it could become with the reality that leasing is still an emotional people driven process. Where and how do we strike that balance of being AI forward and AI first, but also knowing that this is a really human-centric business? Yeah. It's definitely a very specific and very tough problem because it is, it's just so cool. Any time a new stuff comes out. I mean there is the recent AI Dev Day where they're coming out with a new agent builder and it's a huge game changer. Various integrations, MCP, all of these buzzwords, I'm even throwing out now. They are very exciting and they do show a lot of opportunity. It's very much, you know, a similar answer where we have to consistently remind ourselves and our customers that AI, it's a tool, it's a very useful tool and we definitely want to pull those things in. When investment in them is going to actually return for them where it makes sense. We try to just have a lot of very transparent conversations about yes, these 10 new things are awesome. These two probably make the most sense for us and we need to do these other three things to actually get it to drive value for you. So it's that and then it's also a lot of product talk that we do all the time anyways, even outside of AI. Sure, we could spend a lot of time doing this thing to basically get you a fun flashy AI tool or we could finish this other feature that is probably going to drive a lot more actual return. We try and empower our customers as much as possible by helping them understand what is truly available to them, what is actually applicable and being a little bit that go between for them because it is, it's a ton of information. So when you're thinking about, hey, yes, we can develop anything but we can't develop everything because time is a bit constraint. And when you think about impact and where AI would be more impactful than team members and vice versa, how do you think about handoff moments? And when the AI knows it's time for a human team member to step in and signaling that it's time for them to step out and hand this off to somebody. Yeah, I think this is one of the most important things that we have to decide over and over every time we are building a new workflow or new interaction. And it's probably the most important thing that we are doing in a handoff. You were essentially trying to take a step back and act like you were never there in the first place. You don't want to be creating more work by handing off. Knowing when to silently take a step back is how we want to handle it in almost every single case. As we get to see more traffic over time and review a lot more real life interactions with how renters actually want to be handled and what is truly sensitive, then we start to just tune over time. It's constant feedback, constant iteration loops. We are constantly learning and improving to then once we've had something out for a couple of years, we can be pretty specific. We know which of those to a pretty specific degree are okay for us to take over, which ones an agent are going to want to commit to. So it's kind of an art, but it's really more just. It's something you have to learn a little bit new every single time when you learn how to handle things. But of course, similar to what I mentioned before, this is a place where we really want to hear a lot from the people who are currently handling those flows. How do they want to handle it? When do they want to bring in a supervisor? When do they need another team to handle it? - I love that growth mindset. I guess it would be the buzz word. We would call it is it can always get better and we can always fine tune it. And I also love that it considers the customer experience because I think we've all had phone calls where we've yelled into the phone at a voice AI, representative, representative. And we've also had phone calls where a voice AI was able to get us very quickly to the right person. And one of those is a great customer experience and the other is super frustrating. Faster, we can get people when they've given us signals to where they want to go, the better off we all are. So you've talked about the research that we do and how we always go back to our end users and try and figure out and verify and make sure that what we're building is truly helping empower their days and making their lives easier. Can you tell me about a story or a surprising way that you've seen leasing associates use AI in the wild and it made you think, oh wow, that's how we're changing their life or that's how we're really aiding their job? - Yeah, I have one that's maybe not the most flashy but it comes to mind first and I think is at least more realistic scenario. So we rolled out a while back suggested messaging. So a little bit less guard railed suggestions to put into the hands of our users, similar to what we were just talking about. We wanna make sure we are on the side of caution if we're doing anything for you but if we're suggesting stuff, you give them a little bit more leeway and we, this is one of those things. We put out there pretty quickly because we wanted to see how people used it when they were sending it, when they were editing it. It was kind of funny because we have our brand voice that we can switch up a little bit to activate differently 'cause people have different preferences for that. But we also knew, as we've mentioned, that again, leasing is a relationship building process whether it's one off or prospect or over time for a resident. And so the way that leasing ages will read these and change how they're leaving blocks in and the things that they are editing were funny because some of it was, we could totally do that for you and that was such a funny thing that we didn't think about that a human was gonna do because we were a little bit one track mind of, okay, we wanna give them the three answers to their question. We wanted them to have the information so they don't have to go search for it. But then they're writing these, it's actually quite nice to see how they like to communicate with people that they're building relationship with and how nice they are about it. Those fun pleasantries that are a little bit more contextual is very fun kind of seeing those on the micro scale. What are people truly doing with what you're doing today? - I think one of our partners, at least at forum, if I'm remembering correctly, I think 10% of their messages were the message suggestions that weren't edited. Switching gears, we have a partnership with Sierra, AI. They have deep roots in AI specifically. Funnel has deep roots in multi-family specifically. How do you see those strengths coming together to create something bigger than either of us could have done on our own? - I am, first of all, I am so incredibly proud and amazed looking back at what our small AI team has built on our own and what we've competed with considering the number of people that we've had on there. But at the same time, the world has never been moving faster in tech and there's a lot of places where we can harness help from other experts. That's exactly where we're leaning in with Sierra, their agent SDK and what they offer for building agents and how quickly they can get set up and how they're able to do that cross industry. And then like you said, we are the depth and multi-family. Even more excited about the things that we'll be able to accomplish together with that and just the speed at which we're able to get new features out and improvements out in that, you know, what our customers are excited and happy to adopt. Like, I think partnership is going to be so important. We are very much leaning into their incredible technological strengths. We bring them the requirements that you need. Here's the potholes you don't want to run into. And then that brings us together with a product we're all very proud of. - I agree. One of the sentiments about them, I don't know if it's their tagline as a marketer, I probably should know that, is that they provide the world's best AI driven customer experience. One thing is you need a really deep system of record in order to take advantage of a lot of the agentic workflows that are possible with customer experience facing AI, which is exactly where your work interfaces. If you had to debunk one myth about AI and multi-family, what would it be? - I think, so there's a ton of AI solutions for various things out there everywhere. And in multi-family, it's almost incredible how many tools there are for various pieces of the workflow, which I think is awesome and that there's so much innovation going on. But I think that one thing that I see the most out of anything, self-learning of AI tools, it is a buzz word and it has thrown around a lot. And we know that it is not used necessarily. And I hesitate to say truthfully, but maybe as truly standalone self-learning as a lot of places like to say that they are. And a face value again, that's one of those things you're like, I don't have to update my knowledge base. I don't have to teach it new answers. It's just going to learn then that's incredible. Which generally, most tools do not actually have that. There's going to be human in the loop. There's going to be some sort of review. There's going to be at things you need to add in. So it's a little bit extra marketing there. But then the second thing is, you don't actually want that to be happening because if you are just trusting AI to think that it can't answer a question, come up with the answer itself and then set that as its own memory and potentially overwrite information that you have given it. That is a recipe for disaster with where we are at. So it's one of those ones that we end up talking about this a lot as we get customers deeper, deeper into testing. They're like, oh, when's it going to learn that itself? And I'm like, do you actually want it to just go and read conversations? And guess, I don't think that you-- and we'll sometimes show them that. We'll do analysis. Here's the potential answers it could have pulled. And none of them are right. And that's not to say that the way that your AI's answering is not going to be right. It is going to be right because it's using info that is verified by you and by us. Adding in this self-knowledge transfer generally doesn't happen because of that. And also, it's not something that we want to be pushing into as much as you think. This is another just great example of. Humans are always going to be necessary. You always want human with human sense in the loop, even if it's a light touch of review of, OK, thumbs up to this piece of information. I always want a human to be there. And I don't want to tell you that there's not a human there because I'd be afraid. That's a great distinction. I think that's something that we talked a lot about prior to our partnership with Sierra when we were building these models internally-- Chachy-P-T, right? To use an example that probably a lot of us interface with day-to-day or at least once a week is part of the reason new miles are released as the model is being trained and it's being trained, yeah, by AI in some ways, but also, yeah, by people who are making sure it does the things they want it to do better than the last one. And that's not a magical step. And so thank you for bringing that up. I think it's really important with your background and finance and now multifamily and how regulated our industry is. It's important to know that there are some guardrails in place to use this really powerful technology in a really thoughtful, productive way. Absolutely. And you'll even see the industry that big guy is doing a lot of this. A lot of you who are listening to this have probably talked to Chachy-P-T, and you've probably gotten one of those responses, where it'll send you two options. And it's which one do you like better? When you choose one, you are being the human in the loop. That is an example of reinforcement learning with the human feedback in the wild. And you'll start to see that word coming up more and more because even open AI, even anthropic, all of these model providers, they know how important that is. And so they are investing more and more into those sorts of things and also building those into now with the agent developer kits that they're putting out. They're now making it easier for us as builders to do that. And you, if you want to build your own side project to do that, because it is so important. So they're even, they're leaning even more into that. That's a good thing behind the curtain. We're really switching gears here. What is something that you love to do outside of work? And how's that bleed over into who you are and how you show up here at funnel? I'm starting this off on a little bit of a weird foot, but one of the things I love most about my job is that I, yes, I am doing some mundane things in some very technical things, but where product sits, I'm not a designer, but I do get to be very fun and creative when we're left to our element of, okay, you actually have space to go and decide what goes to be on the road map. Go and figure out what needs to be to get done, getting to roll my sleeves up and research and learn. And I am a huge research, something until it's dead person. I'm going on a trip to Mexico City and I am very much in charge of my friend group doing the research and doing the antennae and finding stuff out and connecting this together. And I tried to build a web scraper. It'll pull all of the threads on Reddit for the past year, still everybody who's visited. I love to just dig into these details in that way. And that's something that I also can, and have to do, all the time at work, I really have to spend hours digging into this problem. And I kind of love it 'cause I get to get lost in that world of just learning about what's going on and coming up with something and it's put together a little presentation for my friends. And this is what I have and I'm so excited to present that. So it's a little different, but parallels are definitely there. - I love that. And I picture you on a Friday night presenting, "Hey, we're going to go to Frieda's house, which I hope you do." - I'm going to make my slide with explain my job to a five-year-old, and that's the next one. - I'm here for it. - Valerie, this was really fun. Thank you for coming on. Thank you for sharing your perspective. We really appreciate it. - Thank you so much for having me, it's a great time. - Thanks for listening. If conversations like this are your thing, don't miss the next one live. Forum 2026, registration is officially open. Join hundreds of senior owners and operators at the industry's best conference on AI and automation, centralization, and the human elements of multi-family. Grab your spot now at the link in the show notes. We can't wait to see you there. (upbeat music) (upbeat music)