Will Bodewes - CEO of Phonely | PhD AI researcher

Welcome to another Wise Wisdom pod. We have, uh, Will Baudewijse with us today, CEO at Foundly, PhD AI researcher. Last time we spoke, talk about, uh, his crazy mountain bike adventures and everything that he's been up to. Uh, yeah, Will, how's, uh, Foundly doing? How are you doing on your side?

Yeah, and things have been going really well. We just raised our Series A and we've been really growing out the team, scaling it a lot. We've been seeing a ton of attraction— attraction— traction in the market and a lot of adoption. So it's been a really, really good time for us.

$16 million Series A, that's a pretty fat Series A, probably one of the biggest ones in your space. What are you going to do with that $16 million?

Yeah, good question. We are aggressively hiring in a lot of different areas. I think for us it's a lot on the growth side. So we've had really good product-market fit. We really took the advice of build a product that a few people love and we've differentiated ourselves in the space there and now we're just scaling it out and hiring the team to become top of mind for a lot of different businesses.

Right. So there's been a lot of AI phone companies in the latest YC batch, you're the leading one, I guess, right now. What have you done differently from the the rest of the pack?

Yeah, realistically, we've been working on it for a long time. I think that's something that people discount is, you know, we've been working on this for like 2, 2 and a half years, which maybe doesn't sound like a long time to a lot of people, but for the voice AI space, like we started working on it before it was even called voice AI, right? That wasn't a term. And so we have had a lot of time to build and that means that we've had the ability to solve more of the complicated problems than other people are solving. So where a lot of our competitors are building, you know, chatbots that talk, um, we've been able to build that 2 years ago, and now we build AI agents that are optimized better than any other agent. So when customers use us, they get better outcomes for their customers. That's statistically shown. So example, if you're trying to book appointments with voice AI, we will outperform any other voice AI on the market because we've got A/B testing, automatic improvement. You know, we work really well on latency, stability, all that type of stuff.

Is that due to your AI background? Like, who's in charge of building the full backend? And I guess it's not just a Neil Levin Labs integration at this point.

Yeah, definitely very different than that. Yeah, we've got a pretty good-sized team of engineers that works on that, everything from voice infrastructure all the way to, you know, machine learning, model training, stuff like that. We've started working on and building our own models, which has been really fun and helping us like get lower latency, but also just like have better conversations.

What's your day-to-day as a CEO nowadays? Is it mostly recruiting, mostly hopping on big new client sales calls? What does it look like?

For me, it's usually a lot of recruiting. I was doing most of the sales up until this point. And I'm very happy to have starting to build out like a, a really solid team of account executives to help bring some of those deals in, push 'em across the line, and then myself just jump in whenever, whenever is needed. My focus right now is just like, yeah, recruiting growth, putting the right people in the right places, which is a lot easier said than done for sure.

Right. And what are the top roles you're recruiting for nowadays?

Yep. We're hiring for account executives. It's a big focus for us. We're looking for also people on the growth and the marketing side. So if you're somebody that is looking for a position and is really good at capturing attention, we want to talk to you. We're also hiring for a few engineering roles. We have quite a few folks in late stage on some of these, but we're definitely always hiring. And then we have a fun role, which is like bring your own role, which is we look for talented people that maybe we don't have a positioning opening for them yet, but they we can find some place to make them fit in what we're building.

Tell me about your dogfood strategy. So do you use Phone Leads to take inbound calls? Do you do some outbound campaigns as well to sign up new clients?

We don't do outbound campaigns just because of the regulations around that. So outbound calling has significant legal implications to it where you can get in a lot of trouble, especially in the US. But for inbound, yes, of course, obviously. We have a button on our website you can press, you can talk to PhoneLy, and that's how people learn about the product as one way. The other way that we do is we do a lot of self-serve signup for our customers. So if you're a business and you're looking to adopt PhoneLy, we will help you build it out. And then we use that information to improve the product experience going forward.

So if there's serious legal implications, you push the outbound call product to companies outside the US or what?

Um, not necessarily. So there's a specific compliance standard that you have to follow for that. And so we're not going to be the ones that's like uploading a lead list and calling a bunch of data, cold calling them to use our product. Realistically, we— if a lot of customers, example, like appointment reminders, if you have that in your type of business, placing up on call is a good use case of that. But Since that's not really how we get our leads, then we don't really use our outbound calling product that much.

For the fundraise, I guess you were the main person accountable for that $16 million, which is pretty impressive. How did you end up raising so much and so fast?

Yeah, yeah. I think a lot of it was we just always focused on building a really good product, building relationships, and, you know, our customers had really, really good things to say about us. Actually, 4 of our customers invested in the round. Because they believed in the product and what we were building so much. Um, and so I think a lot of it, what it came down to was, um, our customer— our investors looked at the traction that we had, the growth, not only from a revenue standpoint but from a usage standpoint. And then they talked to our customers and then saw that they wanted to invest as well. And so I think that they realized that there's really something here, right?

So strategic partners, sort of. Um, what are fundraising tips for other YC founders that are raising? Is it just building a great product, having the revenue to back it up, some urgency, some FOMO, some deadline, like any tips?

Depends on what stage you're at. I mean, realistically, I'm always a fan of like, build a good product that people like and people love, you know, like realistically, you can get all in the fancy, like, oh, here's your strategy, this strategy, that, right? But it's like, build something that people actually like to use. And when you have something that people like to use and is growing and you get to that point, fundraising is not a challenge, right? Fundraising is something that just comes naturally. And so I think that that doesn't mean necessarily, oh, you need to spend all your days writing code and, you know, being in the product and not listening to users. Like, the real thing is like, talk to your users, see what they have. If you don't have any users, go find some users. And then understand what they like and then build a product that they really love to use and that they're happy to pay more. So a lot of our customers will start on a small contract. Like we have customers that expanded 20x, 30x, 40x, you know, just because of the way that the product works. So that's, that's always my advice. It's like build something that's really good. There's no shortcuts. It's not like raising money is like your way out. Raising money is just, it's just a stair step in, in the path to where you want to go.

Got it. As a guy that, uh, you have studied AI and I don't know, could be making $2 to $5 million at OpenAI or Anthropic, like, uh, don't, don't you feel that like PhoneLeads, uh, sort of below you, like building, uh, campaigns with AI calls vs. like working on frontier AI stuff?

No, I don't.

Why?

Um, I think that the opportunity for what we're building is significantly greater. I mean, yeah, like, sure, if you financially—

sure, but yet, like, intellectually as well. Like, uh, isn't your brain like a moonshot cracker?

I think the— I think the value in what I'm doing now is I get to have a lot of say and a lot of control into the way that a certain piece of information gets distributed in the world. And for me, that is an incredibly valuable and fulfilling experience. And on the top of the fact that like we are building our own models, like I do get to be intellectually stimulated. I do, and I do get to understand the real problem and not just like be stuck in some AI research lab, understand, like listening to what somebody else does. I get to be on the forefront and I get to decide and steer the ownership in the direction. And that's what I wanted to do. Like that's why I'm here is like I wanted to solve real problems for real businesses. And I feel like this is the way in which I can be the closest to the forefront of actually doing that.

You have a master plan to do something with all the data that you're accumulating?

Yeah, like I said, I think a couple times already, is that we're, you know, we are using some of obviously the opt-in from our customers to be able to build our own models.

And building your own models, um, what would that mean for the end customer?

The biggest problem right now is in voice is the LLM is where 90% of the problem is. It's where 70% of latency is and it's where 60% of the cost is, right? So it's all around the LLM and we spent most of our time figuring out ways to create solutions to that, to get these AI models to talk more like people. to follow instructions more accurately and to respond quickly. We spent a lot of time on that. And so for us, what we see is like, if we can come up with a model that solves all 3 of those problems at once, and it's maybe not a great general model, like it's not gonna be generally good at most things, but if it's really, really good at voice AI, it's gonna be super valuable for the market, right? Because that's a big use case right now. And so that's, that's kind of what we see as the opportunity ahead of us.

What surprising insights have you got from like all this data that you have, um, about EQ humans and how they interact over the phone versus how they interact in real life or via text message or, um, yeah, emails, for example?

I mean, I think that the, the thing that, the thing that I've learned more so than anything else is that like over the phone is a real-time communication pipeline. And so a lot of the methods that these LLMs have been built for are not real-time. Like, you can wait 5 seconds to get the information back. On the phone, you can't wait 5 seconds. You can wait a second and that's it. And so the way that we solve the problem becomes fundamentally different than the way that we're solving, um, in other ways. And so what does that mean? Well, it's like thinking doesn't exist in the phone. Like, you have to asynchronously figure out ways to have models do intelligence, look up tools, do all that stuff. And so that's the thing that people don't realize about what voice AI is, is it's realistically like, yeah, you can create a wrapper and you can maybe schedule an appointment or something like that. But to really get in and to like really solve the problem, you have to solve this problem of like, how do I make my models intelligent? How do I make my models think in real time? And so that's what we try to solve.

And if you crack that, and you're already cracking it, a bunch of enterprises would want to get in on the action, and that's a multi-billion dollar market.

Yeah, absolutely. I think that the software that we build ourselves, like the platform, is the easiest way to build voice agents. So we have that going for us already. But realistically, we kind of feel like— you brought this up in the beginning of the podcast. You said there's a lot of people that are building this, right? And it really bothers us. It really bothers me personally that people are like, oh, you're just doing the same thing as everybody else. Because when we started this, we were the only people that were doing this. Like we were one of two or three companies that were working on the problem this way. And we have a lot of battle scars as to like what works and what doesn't work. And so, you know, I think that like, if you just look at it at face value with like not a lot of context, you just see like the same thing. But like when you really get in and understand and like when, like we're really good for businesses that like, are at scale and really want to solve this problem. We're not looking for small, medium businesses that are like, okay, well, we don't really care how the performance is. We're looking for the type of businesses that really care about that. And so we're building our own models because we understand so much context about this, so much more than most of the other people in the space that we want to say, hey, how's a way that we can differentiate ourselves while also solving our own problems? And that's why we're doing what we're doing.

Yeah, a bunch of rappers in this space. By the way, I come into this with more context than just face value. I tried PhoneMe against a bunch of other apps. I think last time we spoke. I didn't stick for some reason. I owned a call center with 200 employees about 4 years ago or something. I'm deeply bullish on what you're building, for the record though. Have you looked into, like, random tangent, have you looked into Bordy? Uh, do you think that this could be like a model of the future in which you have an AI agent that you can call, it can call other people and, I don't know, generate sales or crack investment checks for you?

I actually haven't looked into Bordy. What are they?

Bordy.ai. It's on WhatsApp right now, so you can chat with it and it's going to call you sometimes. So it's like WhatsApp API. It's a bit confused, the model. It's free, but they offer like a $10K per month package for Bordy to be your CEO, which I think is an extremely weak offer, but I'm very bullish on their technology and the fact that they're building an AI agent with a personality sort of like a human, right?

Mm. Yeah, I don't know. I haven't looked at them. I'd have to get some more context before I could kind of really give a good deep dive on that.

What's next for you this year now that you've raised, you're focused on hiring, you're focusing on sales? I do think you're, yeah, obviously instrumental to the strategy and the roadmap. What's next for you and how do you balance that out with like what your co-founder does?

Yeah, for us, it's, you know, we're looking to bring on a lot of people and figuring out, like, first step is going to be hiring those people, and then the second step is going to be all the problems that arise because we hired all those people, right? So I foresee the next part of my, my year just being like dealing with people problems and sorting all of this stuff out so that the, you know, PhoneMe continues to run as this like smooth machine, right? I feel like for the first part of our journey, we were, hey, like, how do we make this work? Like just kind of scrambling around to make this work. And now we're like, okay, we got it working. How do we make this scale? And so a lot of my time is spent on like, how do I build antifragility in the organization so that the organization continues to strive, you know, without me needing to be around and in control of it everywhere. Does that make sense?

From a first principle perspective, how do you recruit antifragile people? What are your criteria?

Yeah, there's like one, one thing to be said on the, like, on the just general positions that you hire, right? Where it's like, if I have one person in one department and that person gets sick or leaves or whatever, right, that is very fragile because then we have to have myself or somebody else step into that, right? So it's like, okay, well, how do we create teams of people?

So you hire healthy people with good immune systems, you get their data, and That's it.

Yeah. No, but it's like, how do you hire the, how do you hire good people that can wear multiple hats, can do different things? And how do you build an organization that people understand that they can step into different roles? I'm very, very, very, very picky on the people that we hire. There's a huge component of it that's like culture. And part of that's because I believe that, and a certain type, like I have a mental model of the person that I believe does a good job and is what kind of work we're looking for. And I think the other part of it is like, I am going to spend a long time working on this. I want to surround myself with people that inspire me and that I like to be around. And I think that those, those two things become very important to me when we're, when we're hiring.

Any endurance events on the map for this year on the personal level? Do you still have time for this?

We are riding our bike to LA this weekend. So our team is doing a ride to LA from San Francisco. So like 500 miles.

How much?

500 miles. 200 miles.

500. Holy fuck. Uh, any breaks in between or?

We're rotating. So like one person will do a section and then somebody else will do a section. Oh, section. So I think we'll end up doing like probably like 80 miles a person or something like that between 4 or 5 of us. That'd be fun.

4 or 5. Yeah. So I guess that's like a criteria, right? Like check if they have endurance, if they can tolerate pain while being intellectual. Thanks for your time, Will. Where can people find out more about you? Phone.ly AI?

Yeah, phone.ly.ai. You can also check us out on LinkedIn or Instagram, or I think we're getting TikTok set up, so those places. But yeah, probably phone.ly.ai.

That was Will and Charles Cormier.