[0:01]Alright, welcome to the Risk Reversal podcast. I'm Dan Nathan. That's Alex Sherman. He's the co-founder and CEO of Bluefish. Alex, welcome to the podcast. Thank you so much for having me. All right, we met, I want to say a month or two ago. Good friend of mine, Ann Bordetsky. She is at NEA. She's actually been an amazing prolific investor in AI over the last few years. She led your Series A, uh, was that earlier this year? You guys raised 20 million bucks. And, um, uh, she's, she's the best. So she's always, uh, making some very nice introductions to me. Yeah, so I appreciate that. And, um, let's talk about Bluefish. Let's talk about what you guys are doing. Who you are disrupting, who you are serving, and how, I think e-commerce in general is just changing so dramatically. We see a lot of industries being disrupted by, um, AI. We're going to talk a little bit about what it means, uh, Agentic AI, whether we're seeing that as a, a real force and a whole host of different industries, but obviously e-commerce, um, is a big one. And then you also have some interesting data, um, from Black Friday, Small Business Saturday, Cyber Monday, uh, maybe even Giving Tuesday, who knows. A lot going on there. Um, but let's start with Bluefish. Um, you are you are a multiple time founder. Um, and maybe we'll hit a little bit of that on the, on the back side. Maybe some advice you have to founders out there. But this is the one here where it seems like the timing is perfect for the business that you're building. The services that you're providing and what's going on with the world because the ground is moving below every industry's feet. What industry did you go out, uh, and take on and, uh, are you looking to disrupt? Yeah, so, uh, first of all, thank you so much for having me. I'm, I'm super pumped to be here. Uh, and, yeah, a little bit about Bluefish. You can basically think about Bluefish as the marketing engine for the generative internet. And so, simply put, our job is to help the largest brands in the world basically gain visibility and influence into how the major AI providers portray their products and services to consumers. So, for example, when a consumer goes to Chat GPT, or they go to Google Gemini, or AI mode, or Anthropic Claude, or Amazon Rufus, every brand in the world wants to know, are we showing up? How are we described to those consumers? Are we recommended? You can think of it as basically the next chapter of search. And so, our platform gives those marketing teams real-time insight and analytics into how they're showing up. How the models are learning about their brands and how they're portraying their products and services to those consumers, and tools to go and influence those outcomes. So, figuring out what's driving the models. How are the models learning about their brands? What's informing the way that they characterize their products and services and what sort of content can brands build in order to help teach those models more about their products and services. Yeah, so when you talk about the generative internet, like let's take a step back to this old SEO world, right? Like things have changed so quickly. So describe what marketing teams, how they were using search before and how it maybe is iterating right now and where you think it's going in the short term, because obviously your platform is giving these marketing teams tools, right, to do the things that they, to sell more products and services and the like. But it seems like the SEO model got turned off very, very quickly. 100%. So, taking a step back, you know, what happens on the internet that we've seen over the last 20 years, is basically, every five to seven years, you have a new marketing channel that emerges. And it kind of plays out the same way every time. So if you think about the rise of search, the rise of social, the rise of mobile, what happens is there's a new technology that kind of changes the way that consumers interact with the internet. Marketers initially and brands don't really know what to make of it, right? Like remember when mobile came out, marketers were like, no one will ever shop on their phones. And when social came out, marketers were like, what is this thing? But what happens is, as consumer adoption scales, it becomes really clear that that's where the eyeballs are going. And marketers realize that that is sort of a new surface for them to engage and interact with their customer, and you end up with a new marketing channel. That's basically what's happening with AI right now, but it's happened in this really compressed period, where when we started Bluefish two years ago, there were platforms, of course, like Chat GPT and Perplexity that existed. But from a marketer's point of view, from a brand's point of view, they're relatively small, like tens of millions of consumers. But what happened next is that Microsoft launched their AI product, Google launched their AI products, Amazon launched theirs, Meta launched theirs. And in like a six to 12 month period of time, the entire consumer internet became the Agentic internet. And consumers very quickly were trained on this better, faster, smarter way to navigate the internet. To discover products, to comparison shop, to figure out what to buy and where. And so marketers, since then, the largest brands in the world, have basically been rebuilding the way that they think about marketing to catch up to this new shopping journey through AI. And so, we work with search teams, we work with content, brand marketing, PR and Corpcoms, that all want to know, how are the major models describing us to those consumers? Are we showing up? Are we showing up positively, negatively? How are we compared to our competitors? And ultimately, what needs to be true in order for us to win in this new channel? Okay, let's talk about then these e-commerce platforms. We know who they are, right? So there was an article, I think it was in the information. I was talking about Walmart today. I think a couple days ago it was Amazon how, you know, they're blocking some of these agents from coming from Google, right? And basically, I don't understand why they would be blocking. You think that traffic into your e-commerce site or platform would be a good thing. You know, people used to go to like Amazon and search for something they go find it somewhere else, right? You know, that that sort of thing. But why, why do you think Walmart and Amazon are unsure about, uh, e-commerce agents coming to their platform? When you think about it, they must be a disproportionate amount of the e-commerce that, uh, exists in the US here. You think that they'd want as much traffic from these engines as possible. Yeah, 100%. So, the thing that's different about AI versus the deployment of mobile, or social, or search, is that all of those legacy channels, they took three or four years to compose and reach kind of cruising altitude. AI has reached cruising altitude in 12 months. And so, everyone is a little bit on the back foot. So, for example, if you look at Google, Google was forced into the AI space. Because as soon as Chat GPT started to be adopted as quickly as it was, it was an existential threat to them from a search point of view. And so you saw Google race to roll out their AI products. And I mean, it was a little bit of a mess. Like you had kind of the AI overview at the top, you had the 10 blue links at the bottom. As a consumer, you were like, where am I supposed to look? How am I supposed to interact with this?
[7:23]And what's happening now is that we're going through the first holiday shopping season that is almost entirely powered by AI. And everyone is racing to figure out what is that actually mean for us? So, how do consumers go shopping for gifts using Chat GPT? What does that mean for Amazon Rufus? Where do they actually transact? Do they buy in AI? Do they visit a retailer website? And all of the sort of hundreds of millions of dollars of marketing budgets, the billions that have been spent over the last 72 hours to reach those shoppers. Those budgets were built and designed in a previous version of the Internet. They were typically planned six months ago, 12 months ago. And so, I think this has been a really critical referendum for essentially, you know, large retailers, large e-commerce platforms, large brands to figure out what is this new world. And so to your point, I think there are a lot of large platforms that are a little bit on the back foot where they're trying to make sure that they still have access to their customer base, that they can still play a role in informing and influencing that shopper journey. And that ultimately, that transaction happens on their website. Um, but we think of AI as sort of a reshuffling of the deck. That this is sort of uh, you know, a new playing field. And, uh, if you look across the big tech companies, uh, every single one of them is thinking, okay, what is our sort of new role in this new Agentic internet? Do we play in the ad side? Do we play in the search side? Do we play in the B2B enterprise side? Are we a commerce provider? Like, what does all of that mean in an Agentic internet? I think the rules are being rewritten as we speak, uh, and you're seeing that play out in real time.
[9:48]Right. And, you know, you just mentioned, like the Google, it seems like they were obviously defending their moat to some degree, right? They have a disproportionate amount of, you know, search ad revenue, right? And they're like, all right, so if we start putting contextual answers over here, the overviews, right? And people aren't clicking on the links and we're not getting paid, right? And so obviously, you know, we're going to see, you know, some form of ad models on these LLMs at some point. It just seems like all at once Gemini has a different narrative, right? They're doing a good job as far as what the product is, but you have to think there's going to be some cannibalization one way or another. So, when you think about them and Meta really, I think they're like 75, 80% of like online ad revenue between the two companies. Does that sound right? Ish? Right. Yeah. Yeah. Now, this has been a huge business, though, for an Amazon. Their ad business internally, right? Walmart. You know, these are tens of billions of dollars and it's pure margin for them. If you're already on their platform, you know, and and you're seeing these ads and it encourages you to, um, you know, buy or check this out. I mean, like you're getting paid, right, for that. So, how, how, how do you think that, um, what what do you think the evolution of that is? Because obviously, like we just said, we can agree that, you know, there's a huge moat that to some degree Google has to defend, right? Um, but they also have to change quick enough and what does it mean, I guess, for a Walmart and an Amazon because they have these high margin advertising businesses. Yeah. So, you know, when we look at, when we look at a Google, for example, You know, they they have this fundamental math problem that they have to solve. So, they do have the customer. They have the customer base. And I, you know, for the sake of clarity, I don't really buy the like Google Doomers, like death of search, like I, you know, Google is a, is a phenomenal company and they have many talented people that work there. And we're seeing that with, with Gemini. So there's a lot, uh, that, you know, they are capable of figuring out. But the fundamental math problem that they had to resolve was our entire ad business was predicated on keyword 10 blue links. That's how we monetize those consumers. But if a consumer is basically just getting their answer at the top and the sort of AI summary they're getting at the top doesn't really have ads, or even if it does, then we need to make with that summary at the top, what we previously made with the 10 things below it.
[13:03]So, how do we make with one what we were previously making with 10? You have seen that kind of play out on the Google Serb over the last 12 months with a variety of different experiments. Um, and of course, like Google controls their ad business. So, can they increase prices? Can they do things behind the scenes to kind of smooth things out? Of course. But what's happening right now is that Google is trying to figure out how to rebuild their advertising business within this Agentic interface. Amazon has the same problem to solve, because, you know, 12 months ago, you would go to Amazon, you go to shop, go to your sort of, you know, Amazon search, put it in, get your results page.
[14:14]That results page is heavily monetized with a variety of different retail media ad units. That is Amazon's high margin ad business that they use to subsidize infrastructure and logistics and shipping and low costs and everything.
[14:52]It is, it is not a nice to have business, it is the existential kind of driver of their margin. And so right now, Amazon's business is transitioning from that traditional search paradigm into Amazon Rufus. And, I think they've had a, you know, I would say, you know, the same sort of experience that a lot of AI providers have had where they've taken sort of two steps forward and one step back. And, um, and so, uh, that being said, uh, they are figuring it out. And I think you're already starting to see ads show up in Rufus as Amazon is starting to think, okay, what is the right way for us to monetize this new Agentic shopping experience for all of the brands that sell through, uh, you know, Rufus? How do we create more engagement with this AI product? So, for example, when Rufus started, it was like this tiny little chat box that was sort of in the corner. It was like a nice to have, it was an experiment. You're going to see Rufus become more and more the centerpiece. Um, the Walmart CTO last year basically said that over the next few years, they want to replace their entire search engine with an AI agent.
[16:21]Walmart has a habit of saying things like that, uh, you know, three years before they actually do something about it. But, um, I think every large retailer is going to basically be on the same journey where they need to meet the shopper where they are. Give them the sort of AI tools of convenience and then make sure that all of their sort of retail media revenue follows the consumer into that new experience.
[17:01]Yeah, so in market cap terms, if you think about it, since Chat GPT's launch three years ago, you know, you saw the hyperscalers and NVIDIA, obviously, a handful, um, you know, of other, like hardware-ish related sort of, uh, infrastructure plays that have accrued most of the market cap, right? You think about it and so, then there's companies like yours that are two years old and you're probably to some degree being really smart about how you're allocating the capital that you just, because you realize you look out there and you say, we're going to be able to take advantage of a lot of stuff that becomes commoditized, right? So a company like yours is going to be able to benefit as all these other industries kind of figure some of this stuff out, like as you just described there. Um, I'm just curious as you think about it, what are some of the mistakes in 2026? Some of these brands are going to make as they think about, you know, using an engine like yours and what the return on that investment's going to be and where they believe, like, obviously you just used that term, skating where the puck is going. It just feels like we haven't heard too many cautionary tales just yet. Like what are some of the pitfalls you think?
[18:47]Yeah, so, um, so Bluefish just released, uh, our first, uh, Black Friday AI shopping insights report where we basically analyzed, uh, tens of millions of AI responses, um, in response to shopping queries. So, consumers looking to purchase within a particular category. And, um, our engine analyzed all of those responses to basically pick out who are the winners and losers of this Black Friday that is powered by AI. And one of the things that that we saw in this report was that the winners in AI and the losers in AI are sometimes very different than the winners and losers on the traditional internet. Um, that we saw, um, you know, brands and different categories, uh, portray to consumers by Chat GPT and the other AI providers in ways that were fundamentally different than the way that those brands were portrayed on like legacy search. And so, going back to your question, I think the, uh, you know, what you'll see in 2026 is a bit of a space race between the big brands to move quickly and get serious about managing their AI performance.
[20:30]The consumer's already there and now those brands need to catch up. Um, and they need to figure out, okay, if I'm in a commodity category, if I'm neck and neck with, you know, two or three other competitors in a particular product category, I need to make sure that Chat GPT knows why to recommend my product versus theirs to a consumer that's in market. And if the model is learning from our website, I need to make sure that that content is optimized for those models. If it's learning from third-party websites, maybe I need to go do some partnerships with Reddit, or with, you know, Nerdwallet, or another kind of paid editorial blog.
[21:29]I think what you're, what you're seeing again is this sort of reshuffling of the deck. And going into 2026, the winners will be retooling for AI. And I think, you know, the rest will be a bit passengers for the ride, where they're still taking kind of a wait and see approach, and they just will be less in the driver's seat. And our job as Bluefish is to partner with the largest brands in the world to make sure that they are back in the driver's seat with AI. That they are influencing and informing their own destiny in this new channel, um, and can actually make sure that they are represented positively, accurately to those consumers in in AI. Yeah, so year over year, so you have this data now that obviously can be tracking for a very long time. And I'm sure a lot of your customers are going to be really interested in seeing some trends. Um, did you see any retailers that had like a sort of Sydney Sweeney effect? You know what I mean? Like there was something that was kind of unexpected and then all of a sudden like, whoa, something just clicked because this new channel is resonating in a way or just the whole the whole thing coming together in a way they could not have seen two years ago based on their prior marketing spends and how they kind of would attack a period like this.
[23:07]For sure. So, I think what we're, what we're seeing is that the content and data that large language models respond to is not always the same thing that search algorithms responded to. So, um, for example, when we ranked, uh, retailers across apparel, across computers, across home appliances, you know, many different categories, uh, Best Buy came out on top by a, by a long sort of stretch. And the reason is because the LLMs were heavily reliant on Best Buy's, uh, content on their website to inform the way that they described products to shoppers within those categories.
[24:14]And so there was something about the content and copy on their website, on those product detail pages that the models really gravitated toward. And so, anytime there's sort of a learning or an insight like that, What tends to happen is that all the other players within that category look to see, okay, well what did they get right? How do we need to change our approach? What needs to be true in order for the models to pull from our content?
[24:50]And so, yeah, you're seeing, you know, pretty heavy disruption across different categories, um, based on what the models are gravitating toward. And that's kind of what I mean by space race. Where I think the largest brands in the world are realizing that ultimately their job, yes, of course, it's to reach consumers, but now they have a new stakeholder, which are the large language models themselves. That if they want the LLMs and AI providers that are built on top of them to represent them more positively, they need to make sure that those models are getting what they need in terms of content, in terms of data. Um, and so that's what you're going to see in 2026. You know that, you know, AI is sort of non-deterministic and it is hyper-personalized, hyper-individualized. So, you and I could ask the same question of Chat GPT and get completely different responses. Like if you type in, hey, I'm training for a marathon, what are the best sneakers? And I do the same thing. Chat GPT is going to look at all of my history with Open AI, all of the preferences that I've expressed. It's going to look at, you know, demographic details. It's going to do the same thing for you.
[26:30]And it's going to have a much more personalized recommendation to you versus versus for me. And so, one of the things that that's forcing brands to do is really think about like, okay, who is our customer? And when they create that content on their website, on those product detail pages, the days of having kind of one size fits all product detail information to train the models are over, right? Like the average product detail page probably has 300, 400 words on it that are one size fits all for every single type of shopper. But an LLM probably needs a hundred, a thousand times that in order to really figure out, how can it have a conversation with you or with me to explain the merits, the use cases, the benefits of a particular sneaker? And so, one of the things that that we spend a lot of time tracking is how does AI performance differ across different customer personas? And what are the implications from an optimization point of view for those brands? Are they telling the right story to models to enable them to tell sort of a story to one customer segment that meets their needs and a different story to another that meets a completely different set of buying criteria? What do you make of, you know, the Sam Altman memo that came out earlier in the week? Yeah, the code red. Yeah, the code red. And and you know, it's interesting, as as a founder, you know, a couple time founder, I'm I'm sure you've learned a whole heck of a lot, you know what I mean, about the products that you're building and how they're perceived and how they're received and the way you're, um, you know, all your stakeholders feel about that. He knows when he writes a memo like that, it's going out to the world. Right. And he knows that everything he does or said is under, you know, a microscope. When you see a founder like Sam Altman write a memo like that to the world, what does it make you think about what you're building? Yeah, I mean, look, uh, you know, there's there's sort of a like an expression up or out and I feel like that definitely applies to AI. And it's forcing everyone to kind of move at the same speed. And so, you're seeing this with, you know, all of the, the kind of the model companies, where as soon as one sort of, you know, ducks left, they all have to sort of tack at the same in in kind of the same way. That being said, uh, what I will tell you as a, as like a repeat founder, is that whenever you see a whole market prioritize speed over everything else, You just know in the back of your mind that that's creating debt. It's creating this sort of quiet, silent debt that is compounding over time, and at some point, that debt needs to be paid down. And I, you know, look, I work in sort of, you know, Silicon Valley. We we we work in kind of these tech cycles that are fast and furious. And our job is to go build generational companies really, really quickly with the best teams, the best technology, the best investors. But at the same time, you know, if you actually want to build a durable company, you need to think about sort of building the engine behind the business. You need to think about the fundamentals. Um, you know, I would say, you know, in our market, uh, investors tend to swing from like valuing speed and like and like top line growth above all else. And then when the bubble starts to starts to crack a little bit, then all of a sudden investors really care about profitability and efficiency. And so there's this kind of ping-ponging between those paradigms. My job as a CEO and and as a founder of a company and and the leader of a team is to make sure we don't get caught in either of those extremes. That are we moving really fast? Are we building for the future? Of course. AI is happening at lightning speed and it's our job to stay on top of that. Um, but at the same time, our customers are the largest Blue Chip brands in the world. Ultimately, they're reliant on us for stability. Like we're there to be their long-term partner, not this like flash in a pan. Um, and so because Bluefish focuses on being this kind of, you know, marketing platform for the largest enterprises in the world. We need to focus just as much on stability and making sure that no matter what happens in this AI market, whether it's a bubble, whether it's not a bubble, we will be their partner for decades to come. Right. And that's that really changes the way you think about building a business. Right. And it's not one of those things like if there was like, let's say, a recession, for instance, um, you know, companies obviously pull back in their marketing spend, but they can't probably afford in this environment to pull back from services like yours that help them make better decisions, right? And that's that's actually something that, I mean, these different tech cycles will kind of say that this is what they do, but like, especially when it comes to marketing dollars, you know, that that's that's like a different. Help us just think about that really quickly because to me, that's something that if you're building a durable company and a durable service, it should be able to kind of hang in there no matter what the economic environment is.
[32:29]Yeah, totally. I mean, so, uh, you'll remember Mary Meeker, um, renowned analyst and, uh, one of the things that she's she's famous for is sort of her chart that basically compares, uh, time spent, uh, by a consumer in various media channels, relative to ad spend. And anytime there's kind of a delta between where consumers are spending their time and the ad spend, her forecast has always been, hey, that gap is going to be closed, that marketers will kind of move into, into that space. And I I completely agree with that. Uh, and I think what we've seen is that AI has brought the consumer into this new channel. And right now, there's really no paid marketing in AI. Of course, that's going to change. Um, but, uh, you know, what you're going to see over the next few years is marketers moving into this new space to kind of close that gap. I think to your point, um, there is regardless of the economic cycle, you see sort of marketing budgets kind of always exist. Maybe they're sort of scaling up, maybe they're scaling down. But especially in any sort of crisis, regardless of of kind of the severity, brands need sales. Um, no one is going to sort of walk away from their customer, regardless of the economic climate. And so, yeah, we think a lot about, you know, making sure that brands are not just able to kind of, you know, be protected and play defense, regardless of the climate. But they also can pick up opportunities. One of the things that happens anytime that you see a new channel, and this was true of search, it was true of social, it was true of mobile. The early adopters, the the brands that were playing offense, they picked up outsized gains because they moved faster than their competitors. They recognized the opportunity. They weren't just focused on risk mitigation. They were focused on what's the upside? What can we gain percentage points of market share here? Can we win new customers?
[35:25]How do we sort of turn this into a revenue driver for our business? And going back to your question earlier, I think, you know, because we work with, you know, large Blue Chip brands, these are the most sophisticated marketers in the world. They have seen marketing cycles come and go. And they are very sober around, hey, what can we do with this one? How do we, how do we find opportunity for the business? How do we pick up percentage points against our competitors? Um, and they'll do that regardless of cycle. Alex Sherman, I appreciate you coming here. I hope we'll do it again. Um, your insights, uh, obviously about the industry, the industry that you're disrupting, uh, where all of this Agentic AI is going. I'm taking the over in the moon. I know you don't care. I mean, it doesn't matter. Like, but what I'm saying is it's like it's it's not the thing, um, if if I'm not trusting Expedius, um, agent to go plan my trip anytime soon, it doesn't change your life, I'm sure. Yeah. Yeah. Right. So, um, well listen, Alex, I really appreciate you being here. Hope we do it again. Yeah. Thanks so much for having me.
