How a Nontechnical Founder Builds an AI First Company

Amir: On this episode of the show, I have with me Kevin Bradt. He is CEO and founder at CasaPerks, and we're gonna be talking about business architecting AI solutions, building an AI-first company, and doing that while you're a non-technical founder, and having a fairly lean team.

But Kevin, thanks for taking the time.

Kevin: Yeah, thanks for inviting me, and looking forward to the chat.

Amir: All right. CasaPerks, tell us all what you guys do.

Kevin: Sure. Well, CasaPerks, we're one of the, premier loyalty and resident activation platforms in the multifamily industry. So we help, uh, quite simply, we help apartment owners and operators attract, activate, make, uh, residents, you know, engage in the community and, and retain them through renewals

Amir: I'm sure that's a, that's a episode on its own. this one we're gonna be talking about you're building an AI first company.

what, what was the thought? When, when did, when did that pop into your mind that this is solving the right problem for, for you?

Kevin: Sure. so I came out of the hotel industry. I was a corporate officer at Choice Hotels.

One of the things I did there was I rolled out Choice Privileges, which was a global loyalty program to about 1,000 hotels. Then I moved over to the multifamily space. I was CEO of one of the larger student housing companies in the world, and one of our properties was giving cream to the market, and I said, "Well, how are we going to market?"

And everybody was doing the same thing. They were giving, like, away a $1,500 Visa card. So I said, "Do this. Go to the bookstore and buy credits, and then I'll do a marketing campaign to the parents and say, 'If you sign your lease with us, we'll buy their children's books for the semester because we wanna support their education.'"

And that was about a $650 spend, so it was half the amount, and we got four times the amount of leases. And so that is a marketing technique that we learned in loyalty in the hotel space, and I found it quite effective here. And then when I looked around the market, I said, "No one's really doing loyalty at all, much less the way that we were accustomed to how it drives behavior and helps motivate, uh, purchase decisions."

So that was really the impetus of, um, where CasaPerks came from.

Amir: When you kind of cons-consider yourself and, you know, diving into this AI world, I guess, where, where, where do you start? I mean, you have a CTO, we know that, so obviously the, the technical side.

why do we need AI first within this space?

Kevin: Yeah. It's a really interesting question because I'll be honest with you, I was, um, leery about it up, let's say 10 months ago, my CTO and I, because I-I lived through the dot-com and bust era, and, and I said, "Well- Um, you know, let's be cautious about it, but he started it in development, and I saw the, the applications that he was doing.

But what was interesting to me, and, and now I found it fascinating, is that it applies across all departments. And I started using it for, uh, uh, business analytics and the insights around financial data. And so I was able to look at-- I used to have to build corporate strategy groups with headcount. Now, with AI, I can build the tools on my own.

And that's where I started really understanding the, uh, breadth of opportunities that AI afforded us. And so then I went back to the engineering group, and I said, "Look, you guys are engineers of software. Even though I'm a non-technical founder, I'm an engineer of businesses. I put together departments, you know, like marketing, finance, and, and how I construct those is, is kind of corporate engineering."

And, and then I looked at the AI overlay on that and said, "Well, we can redesign a business using this in a whole different and new way, unlike other legacy companies that have to, let's say, let go a lot of people and try to retrofit this into an existing business model." I said, "Let's be open, and let's be, um, just cognizant of the-- our abilities to build something from scratch."

Amir: I am seeing a, a trend, a shift of non-technical founders, obviously the CE-CEO, um, position, and coming in with a lot of subject matter expertise. And, and I, I've asked this question, I mean, I think the answer, uh, is, is clear though, but I am feeling that the age of AI and domain expertise is going to be the difference maker when the IP underneath the hood is reliant on understanding particular aspects of a-

a domain problem you're solving.

Kevin: Yeah. Um, we've thought actually a lot about this. And when you look at-- And I actually, in an interesting lens, I looked at it from like a venture capitalist perspective. And before, um, engineering talent was really, um, if you, if you said, um, let's say a dollar, let's say 50 cents of that was based on engineering, and I think it's diminishing now because of AI's ability to code and ability to kind of, um, make that more commoditized.

But then you say, okay, if you're still investing in a dollar, um, where is the value shift in that chain and how you value companies? It's subject matter expertise. It's customer relationships. It's-- The irony of it is AI, from a growth perspective, is moving it back to the human condition, which is having those relationships and having those customers and, and kind of commoditizing the itself in, in terms of, um, um, IT development.

And that's at least our view on it so far.

Amir: How are you approaching this notion of architecting for AI different than all the other businesses you brought to the table and mentioned before?

Kevin: Yeah, great question. What I'm doing is because you have access to the tools right here, it, it's really around a f- a speed, efficiency, and then, and then staffing plans.

Um, we're very small, right? And we're, we're able to be lean to say, "Look, instead of maybe having six software developers, we can get a-away with four," right? Depending on en-enhanced velocities and things like that. And, uh, what we're, we're looking at, um, you need all the functions, right? But then you don't maybe need as deep of a, of a roster in each function, so that it won't re-replace a function, right?

And we, we, we equate AI to you're a chef in a kitchen. Now you're in a chef w- in, in a Michelin star kitchen. You've got a lot of great more tools, but you're still the chef. You're still the human driving it, making the decisions around it. But then, um, but there's a catch 22 with this, and we're also cognizant of this.

Um, with the, the, um, developing an AI first company and keeping the headcount lean, you're also susceptible to, um, what if people leave? You don't have a deep bench, right? And so you, you've gotta balance that, so retention of your employees becomes even that more critical because if you're relying on, on components and you only have a certain amount of headcount, that existing staff, you need to be really careful on how you hire and make really good, uh, plans for them to stay the long term

Amir: I mean- Mm-hmm ... what have you seen, I guess now that, uh, you guys are probably using agentic within the development process- Mm-hmm

what, what have you seen to kinda help against that? I mean, it's inevitable people will leave, but what are you guys doing to- Yeah ... ensure continuity?

Kevin: Well, um, w- you, you, you-- like for our development team, we have enough specialists that if anyone left, we still-- we have coverage right there. And so we've also built in, um, uh, actually we call it Casa Flow.

It's, it's actually taking an agile development process and then automating it, um, with AI, so like QA functions and things like that. Um, we actually have a re-- And this is interesting because we're talking about the technical aspects, and I just talked to my CTO about this, um, about how we're building a company around AI that then actually embeds in culture and how that a- is actually influencing culture because we developed Casa Flow where there's gates, it's like a train leaving the station, and we've got speakers around the office saying like, "Departing to QA," and stuff like that.

So, so it's an interesting kind of use of AI and, and how now outside of mechanizing this process so that if someone leaves, we still-- it's still working, right? But then also how to have fun with it and how it's creating opportunities for cultural, um, you know, building culture around it as well.

Amir: leading pre-AI and post-AI as, as a, you know, I think in this case a CEO, but just leader in general- Mm-hmm ... does anything change? does leadership stay the same or do you have to start adjusting your leadership style based on what's out, you know, what's happening now?

Kevin: Yeah. Um, that's a great question. The, um, the way I would, I, I would look at it and, uh, Nate, who's my CTO, he calls it decision fatigue a little bit because, um, what, what, what, what happens is you're getting a lot of the tactical, um, um, functions taken, um, and AI is doing it. So it's compressing those times to make decisions, so you're making decisions more rapidly.

Like for me, um, I used to have a certain... You know, you, you would think about a certain amount of development time, and now that's shrunk. You think of a certain amount of sales time. Things are shrinking. I don't think they change your thought process around things, but it changes the gaps between the times of making those decisions.

So, so, um, you, you just have to sharpen that decision muscle even faster because e-everything is, is moving so much at a faster pace that you still have to be as nimble as, as you can during that.

Amir: What, what are you using it as, as the CEO? Where does it kinda sit on a day-to-day?

Kevin: All about reporting for me. Like reporting and insights and analytics. And I actually created a report on my reports. Um, so, um, I... You know, you develop all these triggers for all these reports, and I, and I was like, "You know, I need to create a, a report directory so that I can remember all my triggers," right?

And so, and so, and then I have it bucketed in finance, operations, you know, um, brand management and some other things. And so, uh, it's interesting the way-- We're gonna create a catalog for our, our employees when they come on, and then here's the AI triggers and skills that you can use in your job. Because we've interview-- Uh, part of our process is we actually have, during the interview process, they have to submit an AI project.

We give them a problem, and they have to use AI to solve it, and then they have to present to the team. And so then we'll, we'll create a suite of, of, of skills and triggers that they're gonna use and give that to them as part of their onboarding process.

Amir: does that adjust your, your roadmap or how far in advance you'd like to plan for versus you are planning for?

Kevin: We're definitely pulling forward our product roadmap. Um, a-a-and, and, and so that's where, uh, Nate's like, "Get me tickets, get me, get me features," and stuff like that because we're chew- our, our velocity has inc-cr- increased dramatically.

And, and not only that, but bugs and, and, and issues have correspondingly decreased, right? So it's not like... So we're, we're, we're more forward-looking and not backwards on fixes because, you know, um, the Casa Flow, the thing that we do are, is catching a lot of stuff even before it gets out to the market, and that's all a-automated.

And so we're actually really, uh, excited about the fact that, um, we really haven't tested the, the, the, the total bandwidth of the existing team in terms of what, what they can actually handle. And they're building tools to even make themselves more efficient because, again, we're driving the boat and building the boat at the same time, right?

So like, we don't know like what tools that we're using, you know, four months from now to even further enhance that. So it's, it's been a great opportunity for us to really, uh, learn and shape at the same time.

Amir: When it comes to customers and their expectations, obviously expectations have been, you know, you wait until a feature comes out- Mm-hmm ... et cetera, et cetera. Yeah. Has the notion of AI development, uh, and with the AI, uh, you know, agent- Mm-hmm

within the development product, ha-has it changed how you look at customers, customer requests, the feedback? Yeah. 'Cause it is maybe cheaper, faster, easier just to go, "Hey, let's just, let's throw this idea and see if this customer is adding value to what they were looking at."

Kevin: It's been absolutely transformational for us, and we are winning deals on that specific point because, um, some of our customers that are, uh, um...

Iro-ironically, I thought we'd have to start with smaller comper- uh, customers first and then build to the enterprise customers. We're getting enterprise customers right off the bat. And so, um, and they're saying their legacy, um, um, partners would take six months to build product and it's broken. I say, "We can get it to you in three weeks and it'll work."

And so, um, and so our ability to be incredibly nimble and incredibly customer-centric around these feature requests are, is winning us deals left and right

Amir: are there particular areas where you envision the need for more humans?

where, where would you need to hire if you are gonna have to hire?

Kevin: I overstaff in customer support, and so g- our customer activations group. Because I take a hospitality approach to this.

The way I, uh, I really coach the team, I say, "We're a service company first, and how we deliver service is through the medium of technology." And so, um, and so when you look at, um, our landscape, a lot of times it's, it's, it's, it's sometimes where they, they, they sell a license and then they don't talk to them until renewal.

We have a high touch approach with our customers, with our customer support team. So a lot of bodies will go in there because we wanna make sure that everybody's using the platform to the maximum. We, we host webinars, we host monthly meetings with our customers, and so those cu- key customer touch points.

What, what I look at is what is core and non-core to the customer experience, and where I can lean efficiencies in, in back of the house, in finance, accounting, and stuff like that, that's where I'll get the economies of scale, right? Um, but when it's customer facing, I wanna make sure... And, and if you think about it, then I can over deploy in those areas because I've saved on the other sides.

So I can, I can reorient the business to that front of the office, uh, uh, front of the office, um, touch points and still, um, have a higher profit margin than I would have before

Amir: Do you foresee one day... I mean, I, I don't get to ask this question often 'cause, you know, you, you do have that hospitality background.

Do, do you foresee, you know, as AI and these tools get better, that people will start gravitating towards hospitality being delivered via a non-human, whether it's a AI or whatever it is? I mean, do you, do you think-- Because, and I ask that because, uh, when it comes to hospitality, most people, uh, can't hit, uh, speak to somebody fast enough on whatever service they are.

But, um, we know some of those problems probably could be, you know, solved via whatever the solution is in the back end.

Kevin: Sure. Um, yeah. When you, when-- you'll even look in the hospitality industry, keyless, um, you know, keyless check-in, like, um, um, y-where you can just go in and go directly to your room. So wherever it, it, it makes it efficient for that resident experience, um, absolutely.

But, uh, but you have to have a very sensitive approach to it because you wanna make sure that it only augments that experience by decreasing the, uh, friction that that person wants to achieve with their goal. Now, if they wanna come down and if they, if, if they wanna, if they, they have an issue, um, you know, you want bodies and things like that, where you don't wanna...

Like, I've had IVR nightmares where I've called and I've, I've tried to get through the gauntlet of these IVRs and, and, um, you know, it's very frustrating, right? So, so and that, and that has a halo of-- a negative halo effect across your whole experience with that. So you have to be smart, and this is where, to the point that I was making on the support side, we're doubling down, right?

Because if you're tou-touching support, it's for a reason, right? But then you can make efficiencies around, let's say like in, in the industry, the leasing process, right? Where they're trying to lease and they're filling out paperwork. You don't need a human in that. That can be super fast and efficient, right?

Filtering for, for, for properties and looking around, um, knowing amenities. You can answer simple questions like that. You know, that's where I think there's a lot of economies for that.

Amir: Yeah, it's interesting 'cause, um, everyone's gone through some type of, you know, most people, you know, leasing, buying a home, uh, so much paperwork that goes along with it.

I don't think anyone ever would miss having to go through that level of, uh, paperwork, or if it's automated, we wouldn't appreciate, or we could do it on our own time. changing the way something is delivered is almost as important as everything else we've talked about.

Kevin: Yeah. I'll give you a tactical example for us in terms of that. Um, when people request points from us, you know, property wide... And let's say, um, we had one property that had a flood. And, um, um, we actually created a self-service mechanism within the app that they could actually buy points immediately and then distribute them.

Yeah. Our competitive set couldn't do that. It would take a week or two, and, and we could do that now instantaneously. And what, what, what happened is, let's say 30 residents had to go find a hotel that night. This property was able to give them the, the necessary funding to go prepay as opposed to having to not only are you getting kicked out of your property, but then you, you have to come up with the money to go and, and stay at a hotel now, uh, and then get reimbursed, you know, a month later or so, um, salt in the wound.

Here, they were able to distribute that fast, and that automation was great for us because there was no interaction, they got tremendous value, and we heard about it the next day. You know, that... You know, and, and, and they were happy with us, and we hadn't... We didn't even touch it from a human perspective.

Amir: Fantastic. you guys are ahead of the curve, and it's been exciting to, to chat with you, Kevin.

What's a good way- Cool ... of connecting with you?

Kevin: Uh, great. I'm on LinkedIn. Kevin Bradt, um, you know, on LinkedIn. Um, they can reach out for me there, and I'd be more than happy to chat.

Amir: Kevin, I definitely appreciate your time. Thank you.

Kevin: Thanks a lot for your time today.

Amir: that's it for the episode.

Be back again, different guest, different topic. Until then, two things. One, I think Kevin did a great job of talking to us about, how he's leveraging his background as a leader in other domains within this space and how those experiences have translated in architecting around AI and what that means, and also dealing with the lean team but knowing where he wants to make sure he has enough capacity.

And I'd appreciate it if you could share this with somebody else who could also benefit from the conversation. Also, like, subscribe, comment. Let me know how the show's going for you. Until next time, thank you and goodbye.