How Brands Stay Visible When AI Decides | Profound CEO James Cadwallader

0:00The front door of the internet has changed for the first time ever. meaning that we have relied on Blue Link Search to retrieve information for the last 25 years. And now you can talk to the internet and it talks back to you. Every tech cycle feels exactly the same. Building something that works as a performing product in this space time and time again requires people who've been actually immersed in it very deeply. And simply put, the number of those people at the early beginning of every cycle is very small. When you're a startup, the number one bottleneck is hiring. It's just such a competitive market.

0:31if you're trying to hire the best engineers in the world they can go to any startup in new york or nsf they can go to big tech and get paid inordinate amounts of money in those early stages you're trying to create this kind of rock and a hard place where you're you're basically selling a really hard deal and you're saying hey come and join this early stage startup that's unproven. So yeah, it's a big deal when you hire those initial founding team members.

1:06Welcome to Grit. I'm Jubin, partner at Kleiner Perkins, a show where we go beyond the highlight reel and explore the personal and professional challenges of building history-making companies. Today on the show, we're joined by my friend James, co-founder and CEO of Profound, the AI visibility platform built for the zero-click internet. James has an incredible story, not going to college to start a gold-selling business door-to-door. Today, Profound is one of the fastest-growing companies I've literally ever seen. We also have a special guest, my partner, Ilya Fushman. Enjoy the episode. it's impossible to put your finger on one single thing.

1:44But like, it seems like the market is very, very hungry for something like this. Is that fair? There's this weird paradigm where at the beginning of 2024, when we were working on this, I was like, I must be really stupid because why is the world not talking about this? So we started building it and then seeing the world has now everything's caught up and it's become a very obvious idea. Hence, you know, the nature of a very competitive market. But I think these kinds of markets, if you look at Harvey and Legal or Clay and GoToMarket or, you know, Sierra and Decagon and Customer Support, they are obvious greenfield markets.

2:30and it's almost, I think the prize goes to the most aggressive, basically. And I think we're a team that's very well suited to those types of market dynamics. We try very hard. And can you, like, maybe the 10-second primer and like, what are you doing? In a nutshell, we help companies understand and control how AI perceives and talks about them. So like if I'm whatever, a car company, it used to be the days where I'm like, hey, what's consumer says like, what pickup truck should I buy? They put into Google. Now they put it into ChatGPT or Anthropic and they have much less control over what happens with their brand relative to before.

3:14Yeah, it's the biggest platform shift in the history of the internet. The front door of the internet has changed for the first time ever, meaning that we have relied on Blue Link Search to retrieve information for the last 25 years. And now you can talk to the internet and it talks back to you. And how it talks back when it's mentioning a brand or a product or a service is going to become a ball or is becoming today is already becoming a ballroom level concern for every company on the planet. Yeah. And we've built software that helps companies not just understand how they show up in these new environments, in these responses, but also distribute information that's aimed at retrieval agents rather than humans to increase their visibility in the answers, in the responses.

4:04One of the things that when we were having lunch in New York last time I saw you, that you said to me that's really stuck with me. And this was maybe a year ago when I think it was a little less obvious than it is today. I think you said like, ChatGPT is going to be the stickiest product in human history by far. And I remember thinking like, seems like it makes sense. But like, can you double click on, first of all, do you still feel that way? Do you feel even more strongly that way now? I mean, there was just this period of time, which is made, again, what was obvious now feels, uh, or what was less obvious now feels more obvious.

4:42This is during, when we were having that launch, I think this was closer to that time when there was a lot of yap around, you know, the, the models, uh, the foundational models are commoditized. Like I said, you know, this is just a kind of race to the bottom. And I think while there was this kind of arms, arms race going on, uh, with the foundational models, the miracle that came fruition was chat gpt just won the consumer mind share and yeah it's it's because of the personalization in particular i think the more that you the more you use chat gpt uh the better it knows you the stickier it becomes and i think that that flywheel holds potential for this consumer application to become the stickiest in history it'd be like ditching a friend or something.

5:38Can I go back to the, like, what is happening in the business right now? One of my favorite things, I was actually telling Elia this a couple of weeks ago, one of my favorite things that you do that I think is so funny is every time you hire a new employee, you guys take a ridiculous picture together and then post it on LinkedIn. And it's, you've hired like 80 employees, 90 employees or whatever in less than a year. So like every day there's a new ridiculous picture. What was the genesis of that? And is this, how long is this going to last? Because it's you and your co-founder that are in every picture.

6:13Yeah, I mean, it's something, it was definitely not part of some, you know, master plan. It was just Dylan and I. So Dylan's my co-founder CTO. He's got a good sense of humor as well. Like he's, you know, both of us have fun with it. And I think when we were first hiring, We just wanted to announce someone joining. And I think, you know, when you're a startup, the number one bottleneck is hiring. It's just such a competitive market. If you're trying to hire the creme de la creme, if you're trying to hire the best engineers in the world, which we are, you know, a good engineer, or I should say a great engineer has total optionality.

6:52They can go to any startup in New York or NSF. They can go to big tech and get paid inordinate amounts of money. they could probably come down here to menlo park cough twice and raise five mil

7:07and yeah you know you you're trying to in those early stages you're trying to create a mirror you know there's this there's this kind of rock and a hard place where you're you're basically selling a really hard deal and you're saying hey come and join this early stage startup that's unproven so yeah we it is a big deal when you hire those initial founding team members so it was kind of a an outward celebration of hiring some of the most talented people we would just take a photo with them and you know as we were taking i think as we were taking the photo we thought it'd be funny just like i think the first one was we took a photo someone had someone stand on the desk in the background just without calling it out at all just just uh as a sort of easter egg in the back of the image and then it got that people thought that was funny.

7:55And then we did it again. And it kind of, I think we've had the, it's got to this place now. Yeah. Where sometimes we're hiring, you know, we'll have four or five new people joining a week and we have to take a photo with all of them, but I don't know, I've come, I've kind of gone full circle and I quite enjoy it now. I need office space thing on LinkedIn. Was that a shtick or was that real? like the like where everybody was posting i think you guys got a sign in the new york times or in the in the whatever in the whatever times square didn't you like didn't you guys get like billboards didn't you guys like do like crazy things to get office space at first i was like this can't be real i mean there's sort of two stories here the first is just like the use of these billboard trucks we we have this propensity to using these billboard trucks right now uh they're kind of cheap and they're like trucks with billboards on the side that you can like drive around you pay x number of dollars and they will drive around a location for you we discovered this so we've been using these billboard trucks uh yeah across a lot frequently we've had them outside like competitors user conferences uh we've just been randomly driving around driving around with random messages on them but yeah we silicon alley is is real right now uh high the silicon alley is real in the sense that if you want to find a good office space in new york city and you want to be in the mix you essentially have to be around union square or madison square park those that sort of from 14th street up to maybe 32nd or something in that sort of region and it's just impossibly hard to find good office space around there.

9:42Extremely competitive. So yeah, we hired the billboard truck and had it drive around. It turns out, I mean, if you were looking for a great growth hack to attract brokers, that would be a really good one. We must have had about a thousand brokers message us. The thing that they really won me over because we had our in-person board meeting a little while back is there's this fridge in the profound office. I actually tweeted a picture of but it's basically packed with protein shakes and Celsius. That's the only two things in there. I was like, this is really what a winning startup has in their fridge.

10:15Are people like like you're still in that office today? The original one. We're in our second office now. Okay. But like you've been there for a few months. Yeah, correct. Like, are you on top like on top of each other? Yeah, there's like a what is what is it actually like? Are people sharing desks? no we've had to start deleting meeting rooms and putting desks in meeting rooms it's honestly the reality of it is there's this kind of like giggly moment at first where everyone's like haha this is so funny we're so crammed in and then after a few weeks it starts to get a bit miserable and people it's it's just it's just too much yeah so we definitely need to find a new space um but yeah we're like sardines in there right now it's good vibe everyone's yeah very connected with each other.

11:00It's good for communication. You're expanding though, kind of internationally, regionally. Yeah. It's a big, big moment for you, right? Yeah. We, I think this speaks to a kind of wider topic of sort of AI timelines. I think, you know, what would have been normal in SaaS, uh, timelines, like maybe it would have been three years in, you start thinking about a UK or European office. Um, we're doing that now. So we've, just signed a small lease for an office in london let's just say yeah we just signed a lease for an office in london we've just done the same for an office in sf it'll probably be predominantly go to market focused in london um and then engineering focused in sf yeah this is just so much stuff like uh it's a it's just like i it was a year ago when you presented here not even a year not even a year ago yeah i quite i mean yeah probably closer to six months i mean it was like june oh no yeah i guess when we were here yeah sorry yeah january no it's at the end of february yeah yeah i just so much stuff i don't even know what to make of i just so many so many things for such a young company to be doing it's like uh i don't know it's like you hit puberty at like 10 years old like it's just weird suzanne on my team who you know threw a cmo thing for you and it was like 12 12 CMOs and two months later, she's like, Jubin, it was a breakfast or something.

12:32She's like, Jubin, all 12 of them are customers now. None of them were customers before. And I'm like, what the hell? Like, this just doesn't happen. I mean, why is it such a hair and fire problem for these folks? Yeah, I go back to my point that it's a really, really big platform shift. It was a huge Fortune 100 brand, you know,$50 billion market cap. And essentially the business is like a well-oiled machine. Their gross margins are immaculate. It's run perfectly, except for they have this one reliance, which is all of their top of funnel relies on Google. and it's almost one single point of failure that you know just to even have a glimmer of threat that that could change becomes mission critical for at a board level and they're seeing you know integer level integer percentage referral traffic uh start to come from ai and they're seeing their competitors are showing up more frequently than they would like.

13:44So yeah, it's a boardroom level concern for most brands right now, how they show up. Is that traffic? Like what's the quality of that traffic? Yeah, the traffic's higher quality. Because if you think about the user behavior, most people now have used a platform like ChatGPT to discover a brand or a product or a service that they want to buy. I mean, let me ask you, Jubin, what's the first thing you do if you find something on ChatGPT? What do you mean? Like, click on it. Like, go try and find the thing. But do you click the citation, or do you go back out and go to Google? I certainly don't go then to Google and then research the thing.

14:21Yeah. So I think most people are going direct to the brand. Yeah. So you will go from Chat... All of the... It's essentially become like a dark funnel where all of the research is happening in ChatGPT. So when, by the time that click, takes place, the user has done the majority of the research already. So it's just, it's just transactions happening post-click basically. But eventually that will get folded into the models. Eventually you will just shop via ChatGPT. Like we're already seeing, um, ACP, Agentic Commerce Protocol from OpenAI. We're, um, yeah, seeing AP2 from Google. Like eventually you will just transact inside of ChatGPT.

15:03This is becoming like a consensusly good space. Like it's very obvious that like there's a platform shift that's happening. People have to do something about it. It's obvious to us. It's obvious to the model companies. It's obvious to apparently like public CEOs that want to get into this space. It's obvious to other startups. Like it's very obvious, right um maybe can i pick apart like each of those personas and get both of your reactions to like how you think about them in this space the first one is like i'd be surprised if the models don't do something around here do you guys agree with that or do you disagree with that well i think uh excuse me i think the job of the model is to collapse that that research piece and just be completely trusted by by the consumer right so if i'm like trying to find the best running shoe for I don't know, running a marathon, not that I do that, but, you know, I would sort of go on Chachapiti and say, hey, what's the best running shoe for running a marathon?

16:01And it probably spit out, you know, Nike, it probably spit out something else. And if I'm now, let's say, New Balance or, you know, pick a different brand. And by the way, like as a consumer, I'm kind of lazy. So if it spits out three things in that little window, that's kind of good enough for me. And maybe I'll do a little bit more research. Maybe I want like sturdier heel or i don't know i don't really know what what shoes you use for running a marathon but um i sort of implicitly trust the model and so in some ways the further the time goes on the more i'll trust the model because it'll know me better because of personalization because i will assume it has done kind of the research and summarization of all this kind of research that everybody else did on the internet um and i'll just click on that link um so yes in that sense the model collapses the research funnel.

16:49I guess our thesis here from an investment perspective is if you're that brand that's not in the three results that show up in the little window, how do you get in there? You still need to figure out a way to do that. And that doesn't really go away. Yeah. I mean, and even post-sale, if you think of software as an example, so let's say, let's use MongoDB as an example for the last however many years if you were an engineer that was using MongoDB and you needed to troubleshoot an issue or a problem you'd have gone to docs.mongodb.com now engineers ask cursor as a proxy I guess for the anthropic models or they will ask chat gpt directly and the the big question is okay well when chat gpt is trying to troubleshoot a problem with MongoDB, does it go to docs.mongodb.com or does it go to some funky homemade YouTube video that someone published five years ago that's out of date?

17:49So where do the models go to answer? It's not, it's even the surface area is even bigger than just do you show up in the answer? It's these models opine, they give suggestions, they tailor and personalize they you know it's and i think this is the the misconception around our business is that it's seo for ai i think these models are essentially replacing a lot of the behavior that uh that the end consumer would have done themselves uh which is go and do the research build an opinion create a point of view so really you're as a as a company over the next five to ten years you're going to have to become comfortable with this concept of building marketing for machines instead of humans you're building you're distributing information distributing content and marketing that's aimed at a bot to perceive rather than a human because that doesn't really the differences there are that that type of you know bots don't particularly enjoy storytelling uh they don't care about your opening hook they don't care about the narrative rhythm uh they they just want the information and then the story takes place post ingestion post ingestion it takes place chat gpt creates the narrative chat gpt creates the story so as a brand it's a very different exercise.

19:19I think to your question, will the models release their own analytics at some point and provide some visibility into what's going on? I think, yeah, that it would certainly not be surprising if they did, but I think the market is more fragmented. I think the distinguishing feature of Profound is that it's not just helping you understand what to do, but we actually help you create the content and distribute the content from within our platform. so it really is becoming this workbench for marketers not just to understand how they show up but uh take action and create content so it's really there's two inflections around our business we have the consumer behavior inflection which is people now use ai instead of traditional search and then you have a second inflection which is the technology is now capable of doing a lot of the work that would have previously required teams of humans.

20:11And we're not on, to be clear, I think we're not suggesting that you won't need humans to do this work, but I think humans will be in the loop and just made super effective with our technology. And it's not just humans at the brand, right? It's humans at, you know, there's a ton of agency. Yeah. We work with the biggest agencies in the world. We work with brand teams and they're using Profound today to create content. And it's not slop. It's not a case of we don't believe in one-click content. I don't think that we will ever exist in a place where you just click a button and publish a thousand web pages onto your.com.

20:46I think the human review will be enduring and very important. I think folks who work at agencies, I think SEO teams are very well suited to these workflows, but I think humans will play a very important role in that review. but they'll be more like editors rather than, you know, having to write the words themselves. Um, super interesting. And then Ilya on the, like, I mean, you, you've said this a lot internally, but like there used to be days where the really good categories would have maybe one to two great companies in them. And now we're in the days where there's like seven or eight companies, especially in the early days.

21:23Um, then over time they start to break out. Like let's take Harvey, for example, like started to break out. profound is starting to break out. But now you're seeing like everybody, like you're just, there's a lot of people that are starting to flank this space. Like as you think about it, okay, they're going fast, like check, but they got to go fast. But like, how do you internalize a market that's so obviously great and big, but also everybody else obviously thinks it's great and big. And like what James is doing, like living in this eye of the hurricane of like, he's moving fast. It's impossible to move faster than what they're doing right now.

21:59Like, how do you even, yeah, what do you think? Yeah. And I mean, the interesting thing here is back to kind of the opening here. When we, well, first of all, we met way back at the seed and this was sort of like intuitively made sense, interesting, but it wasn't obviously a huge market, right? Because at that point, I'd say that just even the scale of where ChatGPT was in terms of usage, the sort of proliferation of other agentic interfaces, whether it's perplexity, whether it's Gemini, whether it's what Meta is doing, whether it's, you know, and the kind of the relative traffic from all of those things wasn't as obviously massive as it is today.

22:37And in fact, even at the A, I'd say it was something that you could bet is going to be a really interesting space, but it really became obvious kind of overnight. So we, you know, we did the Series A and Jan Feb of this year. And I'd say by, and the plan was to go kind of, you know, call it one to five. And then by, when did we revise the forecast? It was sort of like April, May. Yeah. Yeah. Revise a forecast like 4X. So it became a step function, obvious space. And I think there it goes back to a lot of what James has been talking about, which is like, what's your unique insight? Kind of, how do you go really deep and understand the needs of the customers, the needs of the consumer?

23:20How do you build a platform that scales? And how do you really do it with incredible momentum? And I think this is where, you know, once you're in the eye of the storm, maintaining that momentum on team, on product velocity, on just scaling, on getting faster to international expansion, on getting, you know, faster. I remember having this, you know, you had kind of an existential crisis in some ways of going from one office in New York City to having a remote distributed team because so much of the build culture is in office to date. But you have to kind of go in and break that glass and then just move really fast.

23:54Can I just paint a counter narrative for you? Which is like, we've never seen companies grow this fast before. Just like never seen it. So like, how does it, like, is there a thing that's too fast? Like you, what is it? Like you lower the quality bar of hiring. Like where do the things of Too Fast start to like deteriorate the business from within? Yeah, I think that Too Fast definitely exists. I think you can certainly break the model if you push too hard. I'm a second time founder. So, and I say this with no hubris, like I certainly don't have everything figured out. But the thing that I've been very particular about is building a really strong leadership team around me so we've brought in people like mark ebert who took six cents who sold to a similar icp all the way up to 250 million dollars uh you know managed hundreds of people in a go-to-market function and now he is leading he's i mean it sounds like i'm just you know bigging up mark but he is recognized as one of the best sales leaders on the east coast probably in the world right now and yeah mark leads i go to market function uh i think uh what dylan's built with our um product engineering team is second to none i think if you want to build beautiful functional product that utilizes the latest ai and you live in new york city or i guess now in sf profound is an incredible place to work because you get exposed to just some of the best builders in the world.

25:34Yeah, our team is very tenacious. We don't put a ton of emphasis on logos or background. We are scrappy. We're super fast moving. But yeah, I think, you know, you have to be careful not to push too hard and break. It's also, I'd say, like, it's pretty easy to look at this and say, well, it's just, you know, anybody can put out a free product that, you know, figures out where search traffic is coming from through agentic systems. And like, sure, like, when you guys were starting out, way back when that was kind of the V1, but so much has been built since then. And there's so much depth and understanding of how the space will actually evolve.

26:10And I think as a team, you know, watching James, Dylan, the crew kind of really understand and formulate the long-term view on this market, I think that's where you kind of have the real advantage. And then staffing, as you said, it's staffing against that on just incredibly high quality talent. And, you know, I think more follow the direction from you guys in terms of the nuances of how you win the space, but just seeing which kind of incredible talent want to join the company this early is probably the best signal that we can have. I think, frankly, every great product starts looking a little bit like a toy.

26:43I mean, look at Dropbox, right? Early on Dropbox, people said, well, it's just our sync with a UI on it, right? Like, how hard could that be? But it turns out it's really hard to build a phenomenal sync engine. And it turns out it's incredibly hard to make it work on like every flavor of Windows, every version of Mac OS, make it super seamless in a way that it doesn't ever lose your data. So just even, you know, and today, by the way, the problem is amplified back to your point of, you know, kind of if you look at early stage venture capital investing, you have incredible teams who can whip up a V1 prototype of pretty much anything overnight.

27:17And you have incredible tools to do that. You have all this knowledge that's been accumulated about startup building over the past decades that's now culminated and easily accessible, frankly, in things like ChatGPT and Cursor, right? Because it is really a summarization of all that human knowledge. So building a demo, building a V1 becomes really, really trivial, but the real value in these things isn't going deep. And I think there it's really two things. One is it's understanding where the nuances of the product and where kind of the nooks and crevices of what makes it truly exceptional are.

27:48And two, it's really, especially in AI, it's understanding how to kind of capture the really big market, right? Because all the markets are up for grabs today. All of them are undergoing a transformation. Some of it will happen from the inside, from incumbents. Some of it, a lot of it will come from startups that are coming in from the outside. but it's really just going to be this kind of battle of who breaks out beyond kind of feature and captures the whole market and business transformation. I think for that, you really need to have depth and time in market. And that's kind of what the challenge for from an investment perspective is you want to bet as early as possible, but then you want to see which of the teams actually have that real unique insight and depth and can have the opportunity to kind of win it all.

28:31Can you maybe go one more click on why you think everything feels up for grabs right now? I mean, intuitively, like think of any piece of any piece of software that's been built, whether it's enterprise consumer, whether it's in verticals, like, you know, kind of more regulated verticals, like healthcare, finance, legal, just we've kind of been historically constrained by the UI, UX. Everything had to work in some set of forms, some set of fields, some set of inputs and outputs that are either human or computer readable. Everything had to be kind of a separate app. Folks had to have multiple apps that kind of talk to each other, and there's a lot of plumbing around that.

29:14But the sort of future beautiful eventuality is you go to one interface that may be voice, that may be written, that may be something else, and you tell it what you want it to do and then it just does it. That's obviously the long, very long dream to dream. But in that scenario, literally everything can be rebuilt. You don't need very complicated UI or X unless you're doing kind of data analysis. All these things can talk to each other on the back end. Every piece of software that's been built in the past could be rebuilt without having to constrain itself to kind of the old scaffolding. We're going to have a plethora of new interfaces, whether it's glasses, whether it's something else.

29:54In the future, obviously, mobile and screens are winning great and beautiful today, and they'll continue to be probably the biggest point of interface. But voice and audio is going to be important as well. So it feels like everything can be reimagined. Now, not everything has to be reimagined. There's not entirely value in just reimagining things for the sake of reimagining them. But you can imagine a world where, hey, you just chat with a thing, and the thing does it for you. And you don't really care what that software looks like on the backend. Yeah. And then that opens up a really interesting playing field for, again, startups and incumbents to kind of reinvent themselves and then capture a broader opportunity.

30:34It seems to me like as we get into the enterprise workflows, call it like wave three of LLMs, if wave one was the models and cursor and wave two was the Harveys and et cetera, and you go into deeper enterprise and into this wave, like the beauty, the magic may be that the LLM can abstract away so much of the dirty work on the backend that can go reach into all of these enterprise systems, all of the different integrations that you have to do, all of the kind of like hair on fire things that we never thought were possible before. Like you can actually like, like there was a transition where enterprise software really, really sucked.