Rebuilding from Inside: How John Waldmann Led an AI Shift Without Breaking His Team

0:00I talked to her about what we want to do from an AI standpoint. and like right out of the gate, I was like, I want to talk to you about AI. And her initial response was, I hate AI. I hate the companies building with AI. Companies building with AI don't care about humans. And it was like a fundamental, it was like I'm coming from a place of living in hospitality. Everything I do is about the human interaction, putting an extraordinary amount of weight and pride on that. And that was like a very impactful thing. But I'll tell you by the end of the conversation, it was a, oh, when can I get access to what you're building?

0:31Hi, I'm John Waldman. I am the founder and CEO of Homebase. We are the leading team management platform for local businesses, serving over 150 ,000 of your favorite local businesses and over 2 million workers. We have spent the last year reinventing ourselves as an AI-first company. Wonderful. Well, maybe one thing that I've really been excited to talk to you about, John, is I know when we reconnected recently, it was because you had posted on LinkedIn about how you are reinventing the business from the ground up with AI. And so I thought it could be fun just to kind of hear you talk, maybe first home-based historically, and then what has changed for you as you've thought about rebuilding the business?

1:19Yeah, I mean, that's a big topic because I think right in this moment of time, it cuts across everything. What are we building? How are we building it the way that we operate as a company? So we can talk about any and all of that. Okay. Well, I like even that distinction, right? The what and the how. What are we building? How are we working? Maybe let's start with the how, because I think that's probably more essential to the DNA, so to speak. And then let's get to the what after that. Yeah, I would say, you know, the what are we building is an evolution. And that part has been very consistent over the 10-year life of our company.

1:56and what I think about forward. You know, the North Star there doesn't really change, but I think that how we build is changing so dramatically. And Jeremy, when I reached out to you, I was sharing some of the thoughts that I was having of like, wow, at this moment in time, I feel like I'm going back to D-school type building and kind of foundational founder mode thinking about product and how to build product as much as I have in the last 10 years. And it was super exciting and fun. in that moment. And I think what I, you know, I'll just, I'll take the way that we build product as an example from like a product and design standpoint, I think getting to the part of like, okay, we're back into the rapid prototyping, working closely with design partners, the pace of iteration and learning, you know, all that stuff has been radically accelerated in the way that we're building any product.

2:49And that's been, that's been one of a hundred ways in which we are, are changing how we work to kind of meet the moment. And when you talk about getting closer to the customer and the speed at which you're prototyping, what's enabling that? Yeah, so getting closer to the customers, you know, I think that's something that we've always been super prided ourself on. You know, we serve local businesses. We serve a lot of local businesses. They're an incredibly fun customer to serve, right? You can go walk into them, talk to them live. Their problems are so tangible. Their stories are incredible. They're the best possible customer that you could hope for you get to a certain level of scale though and like some of those interactions become a little bit abstracted you know you you've got 150 000 plus customers you can launch a a feature in the product and you get hard data back pretty quickly on uh how are people responding did the a-b test hold did it drive the outcome that you wanted you can you can now get survey like launch a survey and get answers back from a thousand plus people so quickly.

3:48You know, we've got 4 million plus people will use home base this year. You can get incredible amounts of data from that. But I think the thing that gets lost in that is like, we're in a moment where we're trying to create magic and like that, that's a little bit more qualitative than, you know, how did it perform on the AB test? And for that, it's like the, okay, let me take a prototype early in how we're thinking about something, put it in front of you in person and see like what are you reacting to you know the the most basic take it back to the roots of building product and design like take it back see what the reaction is come back a week later with a slightly different version of it like a slightly more expanded version and just like keep pulling on that that thread and i think that the great enabler here is all of these new prototyping tools.

4:36You know, we are now probably

4:41pumping out prototypes on Bolt.new or Lovable. Like all of these things have to be so core to the way that we do product and design at Homebase and has really changed the way that we're able to interact with customers and each other. Well, one thing that's fascinating here, John, and I love that you went this direction, is I think when folks who maybe aren't so familiar with collaboration with AI, leveraging AI tools, when they think of AI conceptually, they think it's going to get you farther away from the customer and it's going to put you more towards abstracted A-B tests. But what you just said is now that we're using prototyping tools, we're getting back in front of people more quickly, more rapidly, which I think is a very counterintuitive impact, right?

5:26So what's changed there or why has that led to more in-person almost you could say analog interaction yeah i it's it totally is that analog interaction uh i think part of it is the the pace of development and learning you know jimmy i go back to like one of my favorite frameworks from the d school is that very very basic looks like works like uh you know framework and like are you moving a prototype are you moving a learning or a product on both of those dimensions and those two dimensions are correlated but they're not the same thing and i think in this moment where uh you can be behind the engines working on prompts and working on kind of how the ai is interacting what's possible there at the same time where you're rapidly prototyping on what is the front end of this thing how is this going to show up for users what what are the ways that people want to interact with that i think you just get such a faster pace of learning on that versus hey let's build Figma prototypes that are so high resolution on the what does it look like but are maybe not giving you the same cycle times on just the customer feedback do you take a bolt kind of like prototype or lovable or replet or whatever and you throw it straight to the customer or is that like is that more for testing like do you throw it in production already so we've done all sorts of things like this is the thing we're evolving through in real time uh yeah I sat through a roadmap review meeting last week where it was every single item on the roadmap had a lovable prototype attached to it.

7:03And so instead of talking about like concepts or designs or anything like that, it was, let's talk about, let's just go through all of the prototypes and spend two minutes on each of these prototypes. And these were like out there ideas, right? It allowed us to kind of pressure test the thinking. It's the whole like if a picture is worth a thousand words, then like a prototype is worth a thousand pictures, right? Yeah. I think it was like the joke is true, right? Of like nobody sat here and was like a 20 page PRD doc is the best way to talk about product. And we kind of like ended up there accidentally and probably to the detriment of everybody involved.

7:35But like now you can go back to the prototype and everybody can kind of click on the idea super fast. And that's, sorry, just to be sure, just make sure I understand, John, you're saying that's where we were prior to unleashing our product and development teams with AI. we had kind of devolved, you could say, even though you started, by the way, such a wonderful place, but you kind of devolved to the point that the way we communicate with each other is 20 page PRD docs, which I can imagine even the bandwidth required for a human understanding to get up to speed. I can't get in the field. I got all these PRD docs.

8:07I got to read, right? Exactly. Like, you know, and you think about just like for me as a CEO, then sitting there and being like, Like, well, shit, to stay current on what we might build, that requires so much reading. And the process had become incredibly linear. You know, we proud ourselves and like, our cultural value is no day wasted. That's one of our three cultural values. And yet, like, we had fallen into the trap of very linear kind of product building, very stepwise product building in a way that was leading to these very, very heavy artifacts that were great and served a purpose. Don't get me wrong.

8:44like we got there for a reason, but at the same time was leading to very slow cycle times on feedback and leading to very big checkpoints for us getting customer feedback. And I think the thing that this has allowed us to do is just blow that up and allow anybody to kind of go out and get feedback in a new way. I think as an entrepreneur, obviously you're, the people that you serve are incredible, fascinating because to your point, like these are these incredible people who maybe don't get like the glory and the fast company articles, but like they are the one who keep the economy alive and the American dream alive and a lot of other things.

9:19Where do you think, where are they on AI at this moment in time? That is a good question. I think it's a wide range. Two years ago, I was doing just like a random Zoom catch up with six customers at the time. And I remember at that time being blown away of like, oh, these businesses have moved way faster here than I thought. And I would remember there's this incredible business, Brooklyn Tea, a great tea shop in Brooklyn. And I remember her telling a story super early on. She has an e-commerce site that she runs, too. And she was talking about spending the weekend using ChatGPT to generate a bunch of descriptions for all of the teas that she was going to put on her website and how long that would have taken her before.

10:07And honestly, that was probably like May of 2023. 23. It was relatively early on, maybe even earlier, March of 2023. And she had like found a very specific use case in it. Now, I think of like, there is still a technical gate to like really unlock the power of AI and to push really hard on some of these applications and use cases for which that's very, very hard for a small business owner. And that's, that's where we come in, right, is how are we bringing AI at scale to make building and managing a team incredibly easy so that you don't have to think about all of the minutiae and tasks that come with it.

10:48And I think there, there is an interesting dichotomy. You know, I always remember the conversation with a friend who ran restaurants, very plugged into kind of hospitality scene broadly. And I talked to her about what we want to do from an AI standpoint. And like right out of the gate, I was like, I want to talk to you about AI. And her initial response was, I hate AI. I hate the companies building with AI. Companies building with AI don't care about humans. And it was like a fundamental, it was like I'm coming from a place of living in hospitality. Everything I do is about the human interaction, putting an extraordinary amount of weight and pride on that.

11:29And that was like a very impactful thing. But I'll tell you by the end of the conversation, it was a, oh, when can I get access to what you're building. Because I think as you like talk through it, people are not afraid of AI. They're not, you know, in this context, you're not scared of AI. When you get into the practical applications of, hey, how are you handling no-shows today? Like, do you really need to spend your time on that? Or is that a thing that AI could do for you? Do you really need to be screening the hundred cashier resumes that you got in the past week? Could AI do that for you?

12:05you know you bring it to these very tactical applications it's like oh shit i would love to help actually like i spent hours on that and i hate it uh but you know just connecting it from i hate ai to these hated or dreaded workflows you go oh i didn't know it could do that right yeah yeah and so i think that's a little bit back to our job as product builders is to like bring it back to the real problems that they're dealing with every day and not make it some like philosophical technology, some grand debate around technology and back to like the, hey, we can make your life easier. Could we go back to that conversation?

12:42Because I think that's actually kind of a really cool use case. So you're talking to this hospitality leader who's really, I'm sure, prides herself probably in this kind of handcrafted artisanal care, right, in the hospitality space. I'm familiar with some of those folks. She starts the conversation by saying, I hate AI. What is your approach? I mean, because you're not like an AI evangelist. It's not like you're trying to sell her on AI. But what was the conversation like that moved her in, it seems like, relatively short order from I hate AI to when can I sign up? Yeah, I think moving is like the most trite answer imaginable.

13:24but moving it from the technology to the problem, right? Which is like, okay, let's not talk about AI. What I actually want to talk to you is about how you build and manage your schedule. I want to talk to you about the challenges of hiring. And then with that, introducing solutions into the conversation of, okay, now imagine if home-based at X. And getting those like, oh, that's possible moments. So bringing it back to the things that she's actually dealing with, the things that she's seen firsthand. As opposed to the let's lead with the technology. But I think that's always true. I have a weird question that's kind of like popping up in my brain.

14:03I was talking to somebody else who was kind of talking about the demise of SaaS. And of course, you're doing many like a SaaS product for specific customers. Where do you think the companies like yourself, and I definitely have like a, you know, like a maybe a suggestion. where do you sit in the question of how concerned you should be that a chat dpt becomes so powerful that the uh yeah the features that you provide are kind of like uh just something that people will just do on that app listen i'll be the first person to say i think sass is dead uh i said it And, you know, we had our leadership offsite in May of 2023.

14:50I sat there and like the theme of the presentation was the age of SaaS is over. Here's how we have to think about our company going forward. I totally believe that. I think at the end of the day, if what you are doing is if your advantage is the code base that you had built up over however much time, the size of your code base, the surface area of your product, you are in a bad position because it's pretty obvious that AI tools are going to be really good at writing code and therefore the barrier to entry of just building product is going down incredibly fast, that the moat of a code base is pretty much done, and that the relative ease of expanding our product and things like that is going to come down dramatically.

15:41That's true before you even get to the, hey, what is the role of agents in the future? Just like fundamentally, the cost of building software is coming down. If that is true, then like there's not as much value in software as there was before. So I don't even think you have to believe in agents to believe SaaS is over. Now, if you abstract away from what software is, and then Henry can tell me what your solution here is, I think if you extract away from what software is like well fundamentally this is about helping a specific type of user make their life easier and if you can totally transform that from like hey let's make that work easier by bringing it online designing some relatively easy workflows and building some automations into that to like hey let's use tools to just do the work for you you are a profoundly more valuable company to your customers and that customer hasn't changed you have just evolved the way that you serve them and i think that is the jump that every software company needs to make i don't think every software company is in a great position to do it but uh i believe that we are uh so that's going to be our focus yeah i think i mean like i think you are too which is why i dared asking the question uh i mean i'm biased i have a book coming out on the subject uh august 5th called me my customer in ai and the thesis is basically that entrepreneurship is going to it'd be much easier so you're going to see more types of entrepreneurs to companies are going to look different like so the object that is a company is changed and then thirdly when now more and more people can do it how do you stand out and we talk a lot in the book about this thing we call relationship capital which is you know how do you create the authenticity and the authority to understand a customer so well that you can be predicting what are kind of like the problems that we consult with technology and then exactly what you're doing kind of bundle that up because at the end of the day it requires real cognitive kind of resources to figure out how do you take this specific problem that is in the real world and then how do you apply an lm model to it and another advantage that i think you will have is that while it's not a network effect there's networking this effect in that when you serve so many of the customers that are the same then obviously you see what works and you know like a two-person team you know can't have an innovation team um but you then de facto through the product become kind of like the people that can take those kind of insights and apply it down to to them i i 100 agree with that and one like i listened to the autos uh uh interview you all did and you know i absolutely believe in the thesis around what AI unlocks for entrepreneurship and how great that is going to be for the world.

18:27Totally believe in that. And two, absolutely believe that so much of the long-term value of a company is ultimately in the customer focus and the understanding of a customer. And I think one of the underrated powers of AI is that it, including reducing the cost of building software, is that it allows for verticalization and purpose-built software in a way that it never existed before. And that was true whether you're building software, but I think, or AI agents, but I think it's doubly true going forward. And I think, like, you know, you still end up with these massive pockets of the software economy that are completely horizontal, serving wildly different types of business and users with subpar solutions for all of them.

19:15And I think that is a thing that we should all be optimistic for kind of the flywheel accelerating on of giving more purpose-built solutions. I love you had said a second ago, John, not every company is positioned to make this shift, but we are, you know, from call it your advantage being your code base to your advantage being your customer. Can you talk about how you, because now it sounds like you've been undergoing this transformation for call it the last year. What were some of the key early moves you had to make as a CEO to shift your kind of company mindset from the basis of our advantage is blank to the basis of our advantage is this other thing?

19:53So I think I can sit here and say that, you know, change is really hard. I sit here and say, if like, you know, we were so far ahead in our thinking of this in May of 2023, we're talking about the end of SaaS. I gave that presentation, We talked about it as a leadership group. Guess what? A year later, nothing had changed, like nothing. You know, we kind of continued on our roadmaps. We were shipping the things that we were shipping. We were working on the projects that we had laid out. I took a very bottoms up approach of like, hey, let's let's communicate a vision for where we want to go. But I'm such a I was a big believer of like this changes everything.

20:35This is going to change every part of our product. It's going to change every part of how we work. Let's push this out to the entire company immediately. Wrong strategy in a moment of change, looking backwards. I think where we really started accelerating in that change and getting the flywheel going was actually when we started to be more directive on it. And that was, I don't know, nine months later, coming out and saying of like, okay, actually, we're going to pick some very centralized teams and we are going to

21:07use them to prototype in different ways, whether that's internal operations or the products we're building. We are going to take the people that we have seen the natural curiosity for around these tools, and we are going to elevate them to different positions where that curiosity and kind of experimentation can spread throughout the organization. We're going to recognize that and reward that more publicly in the things that we're doing from a show-and-tell standpoint, from a team learning standpoint, and then ultimately just being far more directive on you have to use these tools you have to change the way that you work and ultimately like this is what the roadmap is and i will tell you as a part of that i took over product again uh so you know i hadn't been running product but i took it over about a year ago for that reason was like we got to be more directive i think it's an incredibly important point you're making and i think a lot of the companies i'm involved in we've seen exactly the same where there's kind of like been this invitation to start with from the founders often and then you know like business as usual uh because you know people don't like to change that much and also thinking like a lot of people who kind of had worked themselves into middle management in various organization are kind of comfortable with the way that they've kind of like learned to do stuff um and so i i definitely hear from a lot of people the same that you're saying is that we had to kind of do that um jeremy and I was talking just before you got on, um, on something that Kevin Kelly was talking about in an earlier podcast where he was comparing artificial intelligence with artificial power.

22:37And then he was talking about how you couldn't just apply artificial power to kind of like the, the wagons that you had, you know, use this horses with. And so you, you had to kind of rebuilt stuff from the ground up on that stack and you had to build factories and with autonomous driving, you probably can't use exactly the same roads and signs. Suddenly you have to rebuild stuff. And so what I hear you say is in many ways the same, that if you want to take full advantage of AI, you kind of have to kind of almost rebuild on a quote unquote AI stack, not technology stack, but kind of like a technology acceptance stack.

23:18Is that what I hear you say? a hundred percent a hundred percent and i i do think um that most people will get to an oh shit moment using these tools and realize how powerful they are and then kind of the natural curiosity can kind of emerge around it but i i do think at an individual level for change you you kind of have to get people help people get to that oh shit moment and then give a lot of the tooling to your point and infrastructure and change and be directive of how different the world is going to be in a post world and i'm sure um you know one of our board members said something really profound uh probably a year ago uh when he was talking about trying to get the company up the up this learning curve and he said is if you want to get people who are uh further ahead on thinking about AI, hire some people that are unemployed right now.

24:18His point was one, they're the people who have the time to tinker. They're not day-to-day in the... They're not steeped in the status quo. They're not steeped in the status quo. They're not busy writing PRD docs. But the other part of it is those people realize that they are not hireable anymore if they don't have a skill set here. And that's a very important thing. And I think if you are a leader in this moment and you care about your team, you care about the people working for you, you're doing a huge disservice to them if you don't kind of push them into the deep end of the pool and nudge them to learn these tasks.

24:57I think there's a lot of people that are kind of doing it by invitation and suggestion, but at some point it's like, yeah, I just feel like I would be doing the team a disservice. Like, this is where the world is going, and if they're not developing the toolkit here, we have not served them. we should be the place where you can get up that. How specifically, so say somebody accepts that premise, John, and they hear you say, I wrote this down. I'm taking diligent notes here. You're doing a disservice to employees who aren't up the curve if you don't push them into the deep end. If a leader accepts that premise and they say, well, where do I push?

25:30How do I push? What are the kind of levers that you use to push people? So there's a lot of, I think, things that are not going to surprise here. It was things like lunch and learns and trainings. It was hackathons. But I'll tell you the thing that made all the difference in the world. We did two hackathons six months, maybe nine months apart. First hackathon was like, hey, we highly encourage you to use AI. Second hackathon was you have to use an AI tool. That is the theme of the hackathon. And we do company-wide hackathons. So it's not just engineering or product design. It's company-wide. Guess which one was more impactful for the change management, right?

26:13It was like, how can I get people over that inertia? A little bit of the fear and all of that because everybody's got like, well, I got all these other ideas that I'm really excited for. And they're not bad ideas. They're always great ideas, but they're not getting people out of the comfort zone in the way that they need to be. And I think part of that was, you know, you start to get the people again that had some of the natural curiosity and were tinkering on it and getting them kind of within those teams, sharing what they're doing and building the agents and showing how they can build prototypes and kind of leveraging some of those bottoms up examples.

26:46Well, tops down stuff of like, you know, take Cogen of like, you will use a Cogen tool and going from like, hey, this is like an optional thing for you to, hey, we expect you to use this as a part of your workflow. So I do think some of it is back to, you know, being a little bit more directive in some of this. And then frankly, it's a lot of the commitment of like bringing the tools in, making the investment of third party tools as well. Yeah. And then how you integrate, because I want to circle back to your board members comment. I think it's so profound and so astute. The comment that hire people who are unemployed right now.

27:23um one of our favorite one of the one of probably the most popular episodes of this show to date is a cmo named russ summers who was a is a very pragmatic episode about how he basically unleashed what he calls a gp team to do the entire marketing function of his new company and we get deeper the conversation we go russ where did your knowledge come from you know what he said i was laid off That's so funny, right? I mean, yeah. And he basically had, and I think one thing, there's two really important aspects to what you're saying. One, somebody who doesn't have a job actually has time to invest.

28:00Whereas somebody who has a job has plausible deniability. Like, when am I supposed to learn this stuff? I'm shipping PRD docs. You want me to stop writing PRD docs, right? It's like this weird chicken and egg problem. But the other thing is that I think is really profound that you said is this person, say Russ or the unemployed that you're hiring, this person knows it's unacceptable. They're no longer fit for the new world and they are approaching their season of unemployment probably differently. So my, so one, I just wanted to affirm what you said there. Two, how did you integrate those people?

28:33I mean, what did you, how did you go about finding the unemployed people who are reskilling themselves? And then how did you integrate them into an environment where you got a bunch of folks who are, you know, some varying degrees of caught up with the status quo? Yeah, I mean, I'm running product right now, so I see it a lot on those front lines. You know, I'm interviewing the PMs coming through as we hire. And I think in the context of an interview and talking to candidates, first, you can see who's curious, who's been exploring the technology, who's spending their time on it, whether they're at a company or not at a company.

29:13and you could unpack the motivations around that, but it almost doesn't matter. It is about the curiosity. And I think we're at a moment in time in a period of profound technology change where like curiosity is rewarded disproportionately. And so I think that is part of it is making sure that we've got those people coming in and folks who have been tinkering and learning with the tools and all of that. I think that's true across every single discipline. you know one of our one of our core operating values is around mastering your craft and like if you're not thinking about how ai can help you in whatever you're doing you're probably not mastering your craft uh there's too much going on so i i'm like a big believer of like a lot of this is top of the funnel and just kind of like the signal through the front door then bringing them in i think it actually kind of happens pretty naturally because i do think that there is like a lot of latent energy people want to learn they want to figure this stuff out but i think this thing to be shown you know they need to get to we just double click on the like to learn this is a question i'm not sure for who are you it might be best um and john you probably from experience and and jeremy probably from you know academia but um i was talking to a friend of mine alice kafferman who was like a you know a high-end coach and i was trying to understand And why is it it's so complicated for some people to kind of like starting to adapt these new tools?

30:44You know, where is it and why is it that some of these kind of like curious kind of I want to learn feelings kind of like why are they not? They always don't. And he was talking about what he called the self-authored mind and the socialized mind. Like basically you had some people kind of like falls into these categories and the majority falls into the socialized mind. Like like to do what we've always done. In an age where you will probably see more and more of our team members do more and more different types of work because like the classic boundaries is kind of like being also torn down by AI.