Experience This! Travel Industry Insights and Opinions
Produced and hosted by three travel industry veterans, Experience This! is a weekly show focused on In-Destination Experiences and how tech and trends are impacting this $280B segment of the world's largest industry (travel and tourism). Part podcast, part YouTube, part LinkedIn, Experience This! is designed for everyone in the sector including small one woman bands up to the largest attractions in the world and the tech, distribution and services companies that make up the industry ecosystem. Featuring Lori Timony, Go City, Christian Watts, Magpie and Bruce Rosard, Arival. Marketing provided by Von Mack Agency
Experience This! Travel Industry Insights and Opinions
AI Search with Scrunch.ai
Show videos are posted on YouTube, subscribe and Like!
Comments and discussions on the Experience This! LinkedIn Page (Follow us and Discuss!)
Podcasts available wherever you get your podcasts:
Spotify, Apple, etc. - Subscribe and rate the shows.
Hey everybody, Welcome to Experience this. We're in Season 3 and this is Show 6, and we've got a great one in store. If you care about AI and AI search, which I think most of you do, Let me introduce our co-hosts and then Christian will introduce our special guests for today Laurie Timoney from GoCity and Christian Watts from Magpie. Say hello to our esteemed audience out there.
Speaker 2:Hey, esteemed audience, laurie Timoney here from GoCity, good to be with you and happy to have Chris joining us today. Looking forward to the chat.
Speaker 3:Hey guys, christian from Magpie, everyone knows I think it's my favorite subject and actually the AI search and discovery part is my favorite part of AI this year, so really looking forward to this. So today we have Chris Andrew, who, I believe, just launched the brand Scrunch AI. Can you just give us a quick intro, chris, and tell us what you're up to?
Speaker 4:Yeah, Bruce, Christian, Laurie, great to be with you. Thanks for having me on the show. I'm thrilled to be here. We launched Scrunch AI out of beta about 10 days ago, early March.
Speaker 4:We've been building the platform for the last 18 months, really working closely with a number of organizations to understand the impact to the customer journey. I think. Very simply, I observed that my behavior as a consumer was changing. I was moving away from a pure Google experience to relying on GPT perplexity, Google AI overviews, to guide my purchases, my research, how I evaluate where to spend time, where to go on trips, and simply, I think that means the customer journey is changing. I'm getting an answer from AI instead of 20 links from Google, and so I'm browsing less. And so we decided to build a marketing platform to help organizations think about well, how do I retain control of my customer journey in a world that I'm kind of getting disintermediated? My brand, my product, my tour, my service is being represented through GPT, Google Perplexity in a different way. So we raised a $4 million round of funding, came out of beta mode in early March and have signed up a little over 25, 30 customers at this stage and growing and excited to be chatting with you all today.
Speaker 1:Let's start with something really basic that you just kind of threw in there. You've been in business, you just launched. For how long ago did you launch?
Speaker 4:We came out of stealth mode late November so we kind of announced beta ending in Feb March. So it's been a little under three months, a little under three months just went public and already raised $4 million.
Speaker 4:How did you do that? You know how do we do that. We've got great relationships from building in the technology space for a long time. I think honestly. I think business is relationship driven, whether you're in travel and tourism or you're in technology or whatever category you're in, and I spent 11 years building my last startup. People tend to move around in technology a lot.
Speaker 4:I spent 11 years building the business.
Speaker 4:I was the first employee to join two founders and built a community of investors and partners that I knew I would want to work with again, and so one of the fun things about rebuilding, building a new company, is finding the people I know I want to work with, investors I know I want to partner with, and so when we were building a new company, I reached out to. Actually, the investor reached out to me from Mayfield, an excellent firm out in Silicon Valley, and said we want to support you guys, we want to partner with you for what you're building, and so it was pretty organic. And then I went out to my network of people that I've invested in and they've invested in me over the years, said, hey, we've got a little bit of availability. Do you want to make angel investments alongside a lead check, and so that process was fairly organic. But I guess it was and it's kind right. It's like you build a network and a community over time and hopefully you find people that want to work with you again well congrats, on that.
Speaker 3:Yeah, so uh you could be bruce. You could be ilia sotskava. He was the chief scientist at open ai and he left and he started a new company. I think on day one it was worth 10 billion. So it's possible to do anything with ai these days there are different levels.
Speaker 4:Good luck he's. He's. I'm not in that science category.
Speaker 1:He's a genius, yeah well, why don't you give us, uh, an overview? You know, maybe start with a little bit of an elevator pitch and then bring it down and and then, uh, maybe we'll do a little show and tell what is scrunch and why should we all care?
Speaker 4:Yeah, I think we talked about maybe maybe I'll just share my screen for a minute for anybody that's doing more than listening and we can flip through just a little bit about what what the customer journey, with the impact of the customer journey, has looked like. And so you know, our thesis at scrunch has been we're moving towards a world of answers instead of links, and I know I mentioned it in the quick overview. But this idea that the first thing we've outsourced at scale to AI without realizing it is browsing, and as we were building the company, we met a lot of organizations that were really thinking about their internal data sets, like how do we organize and clean up our internal data for AI workflows? My observation was that all of the external data in the world was already being used by AI to answer questions about your brand, your city, your product, your experience, anything that's available in the public was being used, and it's used to generate an answer. And so we thought we'd help organizations try to better understand that. And at we thought we'd help organizations try to better understand that.
Speaker 4:And at its simplest, I think we're moving from a world that you talk to a marketer and they're like I used to build websites for humans. Then I started building websites really just for Google, like if you talk to a CMO or a marketing director, like my website is just to get found by Google to show up in listings and maps organic, paid. That shift is moving to large language models. Right, we're moving into a world where, when I ask a question, I get different results from Google organic search versus GPT versus Google's Gemini. And so thinking about what is different, about what these large language models are looking for and the? The quick secret is large language models want language, right they are looking for. And the quick secret is large language models want language, right, they are looking for unstructured data to reflect answers back to consumers.
Speaker 4:And so this is kind of my 30 second cheeky, fun way of describing what this looks like, with a GIF flying across your screen, which is if you've ever Googled for a recipe in the last decade, you get a look at what the SEO optimized internet looks like. This is what we're left with. Right, I Google for a recipe, I click the link. I want to go navigate. I find this recipe site with someone's life story, two videos, an email capture, seven ads, like I can't find the ingredients to the thing I'm trying to cook and why? It's because the incentives are really misaligned. The incentives are for advertising, for cookies, for capture, for tracking, and if you ask this exact same question in GPT, you get a calming experience. You get the answer to your question. You get, very simply, the ingredients, the steps and even when these models are wrong.
Speaker 4:Sometimes I'm seeing consumers habituate to AI search.
Speaker 4:They're just like it's simpler, like whether I'm trying to plan a trip for my family and friends, weigh a hotel for my family of five in this city against this city, or decide the best way to spend a free afternoon in Cleveland.
Speaker 4:These things just give me answers and as they learn me more and more, they tend to refine the answers for who I am right. I almost think we're having a little bit of disruption of word of mouth for the first time. I was on the phone with a CMO of a large brand recently and they said you know, word of mouth has never really been disrupted, but I think we're starting to see that people trust these AI platforms almost as an advisor, as someone that knows them and can guide them, and so at our company we have a theme which is I think there's a lot of fear mongering in the space of SEO is dead. Everything is over. I think SEO is the foundation for how these things are being discovered by AI search. But SEO is evolving and we're seeing that AI is permanently changing the customer journey. So at Scrunch we said, hey, there's an opportunity to help organizations get visibility into this problem, address the challenges and work more directly with the AI crawlers and agents that are trying to get information off of your website.
Speaker 4:So kind of in a two-minute overview that's the landscape that we decided to build in and that's kind of how we're thinking about the opportunity and problem space that's kind of how we're thinking about the opportunity and problem space.
Speaker 2:I think, chris, to your point about you. We sort of started off kind of building websites for SEO and for Google. Then we're building them for. You know, we look at AI overviews now that you know, obviously, part of the Google landscape, looking at chat GPT. So differentiating could be helpful between looking at like an AI overview, which is really part of the Google landscape, but then chat GPT is a separate thing altogether but also one in the same right Because it's utilizing AI. So important to understand, maybe for our audience kind of what it is exactly that you're building for. Are you building for those bots that are scraping and looking as people are using chat GBT or perplexity or whatever? Are you building for the responses in an AI overview or is that strictly related to Google? So maybe differentiating those two might be helpful.
Speaker 4:I think it's a great point. I think the core shift that's happening is the discovery and search process is fragmenting a bit. Right. We for a very long time have relied on Google as kind of the front door. Obviously, there's segments and verticals TripAdvisor is a huge source of traffic and data. There's tour search platforms that are driving volume for folks, but what you just described is now you've got Google, you've got AI overviews powered by Google's AI, by Gemini, you have GPT Perplexity, you have Grok and DeepSeek, you've got meta AI, which is, you know, taking over every search function and every meta property.
Speaker 4:If you've tried to search recently on Instagram, you've probably, frustratingly, been dropped into an AI search rather than searching for your friend's.
Speaker 4:You know Instagram handle, and so I think everyone's seeing that search is where a lot of the money is made and that's where there's an intent to purchase or book, and so we're seeing that to your exact point. Laurie, you need to be thoughtful about what content is being surfaced in each one of these new experiences, whether it's overviews, gpt perplexity, meta or traditional Google and do not bail on Google, right, like your, google foundation is still critical to your business, but we're seeing companies that are tracking referral traffic and citation traffic in these AI models, saying our visitors from AI search are doubling monthly and they're converting as our highest converting channel. And the reason they're converting so highly is because that entire decision process is happening in the native chat experience. It's like I'm planning my trip, I'm evaluating hotel versus hotel. By the time I'm clicking, I'm ready to buy, I'm ready to book, and so you're much further down in that purchase funnel. So people are starting to pay attention to these new channels because the volume is ticking up and the conversion tends to be a bit higher.
Speaker 2:Mm, hmm.
Speaker 3:What do you think of Bruce? Are you skeptical? You like to throw in a skeptical?
Speaker 1:No, I did. I'm not skeptical. I'm wondering how do you know that conversions happening at double the rate, Like that? That's an aha moment here, right? How do you know?
Speaker 4:Yeah, the beauty there is you're seeing a handoff, right, so you're able to see that, ok, this, this visitor came from chat GPT. Ok, that's a new signal in my Google Analytics or my Adobe platform or whatever analytics you've chosen to identify. And so maybe let me back up a step. When we work with companies, there's kind of really three phases of evaluation.
Speaker 4:we start with we take your website your core kind of how people discover what you do, and we help you understand what is the experience of an AI crawler when it goes to your website and tries to make sense of your content, and so you run it through a site audit, we say, okay, when the crawler lands on your website, this is what it understands, this is what it cannot access. But you can see in your web logs when these AI crawlers visit your site. So you have concrete data to say okay, the chat GPT user agent hit my website and is trying to access my information. You can go look at that in your weblogs today. That's kind of step one. Step two is you can actually see the referral traffic now coming from these AI sources so I can see, okay, I got a visitor from chat GPT or perplexity or Gemini.
Speaker 4:Once that visitor's on your site, use your typical funnel conversion channels to say, okay, well, what did this visitor do? And so, if you've got those funnels built out in Google Analytics or the equivalent tracking platform, we are seeing our customers say, whoa, the amount of visitors from these AI referral sources is more than doubling monthly and they're converting at a higher rate again because once they're on your site, you can track them. You knew where they came from and they've done so much research offline, outside of your experience, but hopefully your content was influencing what they were consuming, Because if it wasn't, they're probably headed elsewhere and that's kind of the challenge that new brands and experiences are thinking through. What Bruce is thinking is that's a bit of a red headed elsewhere and that's kind of the challenge that new brands and experiences are thinking through.
Speaker 3:What Bruce is thinking is that's a bit of a red herring and it's all just actually just a different stage of the funnel, like a Google ad versus a Google organic. But that's fine, there's nothing wrong with it. It's like when TripAdvisor came out with their new trip planning thing and everyone converted it 20x. Well, it's the people that are really keen to book something that but that. But that's fine, it is going to keep growing right, and if that ends up averaging out at the same conversion, it's, it's still something we need to have right, wherever they're coming from.
Speaker 1:So well, I love how you know what I'm thinking. That's great christian.
Speaker 3:Um, I just save you from having to say things, bruce, if I can just say it for you but also you could just relax on this discussion yes, converting it double the rate is another right like double.
Speaker 2:That's what I want to know, sorry go ahead, laurie, because no, no, I'm just. I'm just saying, like, converting it, double the rate. Like double, what rate are you referring to specifically?
Speaker 4:No, I think maybe I misspoke.
Speaker 1:Traffic is doubling monthly for most customers coming from AI referral traffic right, no, and okay, that's actually where my question was, and maybe I wrote it down in my notes. So, double from 20 visitors a day to 40 is one thing. Doubling from a million visitors a day to 2 million is another thing. So, like what, when you're looking at these numbers and I'm sure it depends on you know, you've only been in business for a couple of weeks right Like who the clients are? Like what kind of traffic increases? Instead of 2X, 2x of AI traffic? Is that a meaningful number yet, or is that not quite meaningful?
Speaker 4:Great question. We're getting to meaningful numbers, and what I'd say is think back through the history of large language models. The AI web search capabilities really didn't come online until about a year ago. So if we went to GPT last year at this time and asked a question, the data was out of date by six months or a year. Right, it would crawl the website, it would you'd ask GPT a question. Be like warning this information is six months or a year, or sometimes two years, out of date. That was the response a year ago when we asked questions.
Speaker 4:Now, since then, they've all introduced these web search capabilities and so they are accessing real-time data and so you're seeing consumers move to these platforms, because they're getting accurate information in a way that they didn't 12, 15 months ago, and so what I'd say is like we really started seeing these numbers tick up in October of last year because, again, you really weren't seeing much referral traffic. It was only paid users that had access to web search in GPT. Now web search is available to all users, logged in, logged out, paid or not, because search is a way to make money. And so doubling monthly? You're asking the exact right question. Well, a year ago it was zero. I'm watching it double since October, november, starting by absolutely small numbers, but meaningful, right? If you're, let's say, you're booking a cruise and you got 10 visitors in a month and they one person booked. Now you're at 20, now you're at 40, now you're at 80. We're seeing tens of thousands, hundreds of thousands of visitors being referred by AI search.
Speaker 4:Again depending on the size of your overall audience. I think it's safe to say across the board, overall search volume is still dramatically in Google's category. There's a number of reasons for that. Part of it is that you get fewer visitors in an AI search world because they're consuming a lot of information before they visit your website. I think in our largest customers, or the customers that are most mature in the platform, they're seeing the amount of referral traffic from AI move into the double digits. So still not huge numbers, but 10, 20% on the top end of companies that are looking at their overall web traffic and saying 20% of our visitors are now coming from AI. So that's the number we're keeping the closest eye on. So that's the number we're keeping the closest eye on. The speed at which it went from zero to, in some instances, 20, happening in six months that's a trend line that I think companies should be paying attention to.
Speaker 3:Bruce was thinking it's the 20% he wants, that he needs a number like that, which is right, I think. In terms it's almost like we went 2000 with when google launched, to 2010 or 15. I think we're going to get that whole 15 years compressed into the next six months. I see everyone everyone's asking because they're all freaking out about losing 20 of their traffic, plus more to other things, to social and everything else, and everybody's freaking out because they don't know what's next. So it seems like good timing to me.
Speaker 2:So now the question is your business and you've relied on Google for years and years to drive traffic and now you're trying to understand now what do I do Like, how do I shift my behavior? How do I bring in this business from ChatGPT and other LLMs? So that's where you come in right. So maybe you can talk a little bit about what you're doing and specifically I'm thinking of what sets you apart from your competitors. You have competition in this space, so maybe some of the things that you could talk about that are different than your, than your competition and what you're doing to help businesses drive that traffic yeah, who?
Speaker 3:one second? Who are you saying the competitors are just traditional agencies like seo?
Speaker 2:I mean there's other companies that are doing what Scrunch is doing. There's a few of them. I think I wrote down their names, which I'll see if I can find in a minute. There's one Otterly otterlyai, right, I think. They just partnered with SEMrush, which is obviously Google search. Peak is another one. P-e-e-c HubSpot has a, has a some sort of a similar type of, so there's probably about five or six of them anyway.
Speaker 4:Yeah, yeah, yeah, great question, lori. You know I think you'll see more and more companies coming into the category. This is a busy problem area that people are trying to figure out. Our approach is that we're really trying to think through what the marketing platform of the future looks like, given that this customer journey is changing, and so I think I expect that monitoring will largely be a commodity in the space.
Speaker 4:You'll see a bunch of companies where you can monitor how you show up. They have similar ways of getting the information to understand your presence in these places. I think where the differentiation comes in is thinking more holistically about what does the journey look like as someone is arriving at your site, what information they've had access to, and we spend a lot of time building out those capabilities in our platform. So if you visit our site, you'll see kind of themes in the platform that we've built, and we actually start with this idea of personas, this idea that if you know who you are selling to, you need to understand that these people are getting different AI search results from you. Right, if, laurie, you do a search with your experience and your expertise and your background, and I do an AI search, we're going to get different search results because the models have different memory around us.
Speaker 4:Our prompts are going to look different, and so this idea that if you are a travel organization, you have five different types of target buyers we start with a very deep analysis of your brand and a platform that generates and maps personas to the types of people you're selling to. That helps us understand the types of prompts and types of responses and whether your brand is showing up. I think it's a lot less critical to look at generic content and searches across the web and say, okay, well, what are the people that I'm trying to sell to searching and then show what the responses are back from the models. From there you're moving into monitoring and visibility, for how does my brand show up compared to my competitors?
Speaker 4:What third party sources and my website sources are being used by the model to answer questions and then critically moving that into action to say, okay, well, here are the gaps in the areas that I'm not showing up in AI search. And very actionable saying okay, well, this is the type of content that is missing from your site. These are the problems with how you are structuring things. The model arrives at your page and can't retrieve it. So help us address these problems so that you can actually get your content sorted and cited by the model in the results and then see an increase in referral traffic coming through. So I think the core difference is more moving beyond monitoring into insights and then thinking about the layer to better manage and interact with the agents, crawlers and bots that are trying to access your content.
Speaker 2:And is that your knowledge hub? Is that that piece that you're talking about, that kind of maintains and then delivers the insights?
Speaker 4:There's a category of our product called Insights. Maybe I'll just share my screen briefly again. I can bounce back to this. All right, hopefully that's viewable and accessible. I'm just on the Scrunch AI core website. Here You'll see four categories of the product. One is around basic monitoring.
Speaker 4:So this is kind of what I spoke about around defining your personas, understanding your presence, your position and sentiment compared to your competitors across the leading AI models. So am I showing up? How am I showing up and what can I do about that? At the core of this is understanding what sources are being used by the model to answer these questions, right? So is it using my content? Is it using competitive content? Is it using third-party content?
Speaker 4:Critical to remember that really gone are the days of AI just looking at your website to answer a question. It wants an unbiased view, so it's going to go out to the internet, look at third-party, competitive own content and formulate an answer. So our customers are often thinking my brand is no longer what I say. It is my brand experience service is what ChatGPT says. It is what Google AI overviews say. It is because it is now representing my offering to my customer base. So monitoring is kind of the entry point. If you don't know how you're being represented, it's pretty tough to make those changes. So a lot of that mapped out here on the insight side to make this actionable. We help you understand where you're showing up and where you're not showing up and what the content gaps are, what the maybe the title page, title description, metadata gaps are in terms of why is AI not using my content and it's using third-party or competitive content, and so this experience is basically going to show you the gaps in terms of the themes you say you want to be known for, but where you are struggling to show up.
Speaker 4:That often starts with just a simple site audit. We need to understand when a crawler goes to your website, can it make sense of your content or is it so JavaScript heavy? There's been quite a few reports in the leading SEO platform blog saying look, these AI crawlers cannot read JavaScript and so they're skipping your content and looking for answers elsewhere. So making sure you have the right unstructured data, the language for a model to interpret your content and drive answers. So insights is really where you go in to look at the gaps and work on content suggestions to address it.
Speaker 4:You hit on our knowledge hub, which I'll pause after this, is basically saying there's things that you need to keep a timeless eye on, right? So what does my offering cost? Right, if I go to my website, I can probably figure out how much the product or service or experience is, but if I ask chat GPT how much it is to book that tour or to book that hotel, what information is it accessing to answer that question? And we're able to show you that? Look, gpt found 30 websites that you own and the pricing is inconsistent. Maybe you had a blog post with an old offer or an old promo code. It's trying to sort through all this information to accurately reflect your pricing, your description, your hours, your availability, and it's also gonna go look at third-party content, and so the Knowledge Hub is looking at things like pricing. You know, experience, details, reviews, support, location, and bringing that information to light in terms of this is how AI is consistently representing this back. So that's how we think about our Knowledge Hub interesting it's.
Speaker 1:There is a lot there and it's a little hard to for me at least, to grasp it without an example, um, and I don't think you have examples yet of like experiences, company or operators, right, so we need to find some beta users for you um, yeah, and we can publicly share.
Speaker 4:We do have folks in the travel and experience industry using the platform, but I don't have any public case studies I can share at the moment.
Speaker 2:Yeah, maybe GoCity wants to do that, laurie monitoring part and it's interesting to see what it's interesting to see. That number one, the to your point, chat GPT is looking for consistency across multiple platforms, whereas Google is looking for unique content. So you're trying to balance sort of this, needing unique content with consistency, and it's not always an easy place to land all the time. And the other thing that's very challenging is that you know you talked a lot about third parties. Obviously, to people in this particular space, those are the OTAs. So you go to the OTAs and the OTA sites are really built for Google. They're built for SEO. So you know again, you're struggling with the content that you have on these OTA sites, whether or not it's consistent with your own content, while also balancing that whole trick around what type of content are you delivering for? Google right and SEO? So I think we're very much in a time right now when it's hard for businesses to know where to land, and you know we're struggling with that balance ourselves.
Speaker 3:I think, I feel like I've always got the worst internet out of all of us. Wherever I am in the world, I have the worst internet.
Speaker 2:Are you not hearing any of this Christian?
Speaker 3:No, I heard the last bit, but I did drop out for a second.
Speaker 4:You're moving from a world of really just optimizing for Google as the primary driver of content, primary driver of visitors to your presence, to drive conversion to a more fragmented world and like this is changing in real time, like we're here speaking. Last night, gpt introduced more changes to how they're representing search results. I think we can only expect a faster pace of change from the models, from traditional search, from AI search, so it's something that organizations need to be paying attention to.
Speaker 2:So, maybe Sorry, Christian, go ahead.
Speaker 3:No, I just think it's great what you're building. I think it's tough because it is moving so quickly. Right, it's definitely potentially going to get just run right over, but it might not. I, I I actually don't like gpt search um, I, I'd rather use the. It sounds just kind of geeky, but I'd rather use the llm to find information, because I find the llm is far better when it goes to search. It searches the internet. The internet's generally not very good compared to a, compared to the ai, but that's that's my opinion.
Speaker 3:People are using search because they're used to doing search and they think they get a better result when they use the search option. But I wonder if those two come together and it might make tracking really hard, right? If all of this I know Sam Altman in the past has talked about real-time sort of training during back and forth, which might make this stuff kind of not trackable Aside from it just comes from ChatGPT or from Gemini, right? I guess none of us know the answer to that, but it's one of the places it might go, right.
Speaker 4:Yeah, I'm sure they'll keep testing. They probably don't know the answer and they'll. I mean I respect those organizations for, rapidly you know, testing against those ideas right, trying to figure out what the best way to serve that is.
Speaker 4:I mean, deep research is kind of a combination of the two. It's obviously only included in paid plans for now, but if you run a deep research query you know it can take 20 minutes for the answer to come back to you because it's using really the depths of its trained in knowledge. It's going deeper into research sources and pdfs and papers to source an answer for you. So I think that's kind of the combination of modes I think we can expect to see continue to see more of that did you guys see manis yet?
Speaker 3:did you see the demos um yesterday of the itinerary search? It's the bruce, I see it's the latest. It's chinese model and nobody's got access yet. Maybe some people have, but it's. They did a um create five-day itinerary in japan and places to get engaged and all this stuff and it. It went through a whole thing and it was the best I've seen, but it was their demo. Maybe it took 20 shots to get there, who knows, but it the end result looked like it was a single question, full doc, everything you need to plan your trip, and I guess that's what we're going to expect now this stuff kind of moving so fast I think we'll see more verticalized examples.
Speaker 4:I was at a grocery store recently where there was a qr code on the end of the shelf. You can scan it and it would help you find things in the store. I talked to a store clerk and they're like, yeah, we're able to influence what shows up so that we can try to move through inventory, right. So they're basically connecting this localized LLM to their store inventory. Like, well, maybe this product's getting old or we need to move this because of an incentive to backfill something on a shelf space. And so you scan it and it's like, oh, this is what I should cook tonight, or these are the things I should prioritize buying what I'm looking at.
Speaker 4:So I think again, understanding what the incentive is behind the model is kind of important. Like I think the beauty of a lot of this AI search today is it's not easy to directly influence it by gaming it. You can't just make a bunch of tweaks. The model is looking at a bunch of information to answer a question, and so, as we see more localized, verticalized versions of that, I think it's important for consumers to keep in mind. Okay, maybe this one's guiding me a certain direction because of the incentives behind it maybe this one's guiding me a certain direction because of the incentives behind it.
Speaker 1:I'm still confused by what people in our industry can do, even if they hire Scrunch or a company like yours, to really have an impact on AI search right? So I'm Bruce's bicycle tours in Boulder, colorado, and I'm like. I spent money on Google that's worked. I spent money with Meta that works. I use the OTAs. The OTAs are definitely going to have to reevaluate how this all works, since their money all goes to Google. What do you tell little guys like me of how to get into the search on a JetGPT or Perplexity or whatever the favorite search engine of the day is?
Speaker 4:Yeah, it's a great question, I think, for a small business bicycle tours. I would be thinking about two things. I'd be thinking about top of funnel content that surfaces you in AI search results, right.
Speaker 4:So if I'm headed to Boulder, colorado, and I've got a weekend, I would want to see blog content about, like, how to spend a weekend in Boulder the best outdoor activities and if it's only content about your brand, keep in mind that AI is going to try to go get other views. And so I would say, have a blog post on your site that talks about your experience and other ways to spend the day. That is kind of rounded out, and then see if that gets cited in the AI answers and see if you get new referral traffic from it. So content is still the name of the game here, but it's thinking about yeah, it sounds like SEO.
Speaker 3:Yeah, it's the same type of content.
Speaker 4:Yeah, that's what I say. Seo is the foundation for AI search. If you're not surfacing in search, the models are using search to access information. The difference is when GPT finds this page that it wants to look at, it's going to go download your content and look at it differently from how the Google index looks at it. So looking at all the language, downloading it and choosing whether to answer and cite it in the direct result. But I think the core thing is to understand thematically what are people looking for in these platforms. Make sure you have content to match and then track and see if the referral traffic and visits are going up.
Speaker 1:Yeah, cause the links are totally different. Right, I mean, the links are more. If I do a blog post that says all the great things to do in Boulder, it's going to link to like whoever you know, my blog post is as the refer source, not click here to buy a tour. Right, I mean, it's a very different direction.
Speaker 4:And I mean to that point, you're probably rethinking the customer journey once you do have control of that visitor, right? So if they're landing on a blog post, having some calls to action in that post to get them through to book the tour, that looks different because these models haven't introduced advertisement or kind of commercialized yet, so you're not clicking through and paying for that and it's a weird environment, right? Like GPT Sam Altman has said he doesn't really want ads in the models. I don't know if that's going to hold true. Maybe they're going to pay per click and you could imagine microtransactions of every time I send a visitor, regardless of if it's an ad or not, maybe you're paying for something.
Speaker 4:We've been working directly with the Perplexity team, talking with their ads organizations. They're testing a lot of new formats. They don't want to just recycle what AdWords has done over the last 15 or 20 years. So I think we're going to see a lot of testing and experimentation in every vertical to try to keep the consumer experience delightful. It still has a long way to go. It's still very, very early.
Speaker 2:I think I just heard that Google was going to be adding advertisements to the AI overviews. I think that's a new thing that just came up, which is kind of a bummer, because, personally, one of the reasons I like the AI overviews and I like ChatGPT is due to the conversational element about it right.
Speaker 2:So it's not like you're just getting served up all these ads. Likely there's going to be a way of monetizing outside of a subscription. There's going to be a way of monetizing these LLMs as well. You know for for, certainly for chat, gpt, I got to believe down the road there's going to be some kind of monetization.
Speaker 3:Well, I I got into some arguments recently. I'm getting a bit annoyed with the people people who think that your AI agent is going to get bribed. So I think there will. So I think you can put ads on AI summary, but I think it's got to be clear, like it already is with Google. Right, there's got to be a delineator. This is sponsored content and this is. I don't think you can mix the two and I don't think ChatGPT would ever mix the two. I'm just. I know people say, oh, that's crazy. The big company got to make money. But I think the stakes for OpenAI are trillions of dollars. Them selling out to make a few billion dollars. I don't think it's interesting to them. So I think they have to be pure and they have to be trustworthy. And the second we find out we can buy results from ChatGPT, I think they lose their credibility and puff goes there.
Speaker 2:Well, I don't think it's trustworthy now Christian, I mean depending upon what kind of information you're looking for, there's still plenty of inconsistencies and inaccuracies, right now with chat GPT.
Speaker 1:That's incompetence, not that's incompetence, not um corruption well, that's that, that's so they say what's going on in our country right now let's not go there, no interesting.
Speaker 2:so, like understanding where you are in the space I mean, I guess that's one of the things that I was thinking about as a small business that maybe you don't have the money to hire scrunch or somebody else to help you with your strategy you can at least monitor, like where are you showing up in AI overviews? Are you showing up? Are you showing up on chat, gpt, or what businesses are showing up, why you know kind of spending some time looking at it. I would suppose could be a good place For sure.
Speaker 4:I mean, I would just encourage companies to start by looking themselves like go to these models and see if you're showing up Like if you Google, if you GPT for your brand, you're probably going to show up Like if your brand has been real and existed for a period of time. But moving higher up the funnel to, like the shape of searches shifting right, Kind of gone are the days of me going to chat, GPT and typing in Carnival Cruise Like I don't do that.
Speaker 4:Like I might go to Google and type in Carnival Cruise, because it's directing me to the website and I have to pay a bunch of money for my own name. Gpt is much more human. It's much more like the intent of the question of, like I'm thinking of doing something like this, and so if you're a small business listening, I would say go, start by asking questions about your service, your experience, directly in the models, and then move up the funnel a little bit. Ask for questions about the themes of how people might spend time and explore it. Test it in the different models, the themes of how people might spend time and explore it. Test it in the different models. Know, though, that you're getting results that are custom to you.
Speaker 4:Like you might see this, like I've seen GPT-2, this, where, if you ask a question about your business, it'll say XYZ in parentheses your company because it knows it knows it's answering a question about your company because you've asked it, and so you know I'm not getting that same result because I don't own that company, and so you need to understand just how personalized these results are.
Speaker 4:I did want to comment on the shape of commercialization because I think if two years ago or three years ago we'd all been asked do you think you will pay to search online, I think we all would have said no. Like Google's been free forever, it's pretty good at what it does. The number of people that are paying to use perplexity in GPT to have a different search experience is significant. So I think just how different is the future of commerce and transaction going to look? I don't know. I'm trying to keep pace with it and help companies understand their representation and maintain control of that customer journey, but I would not have guessed that I would pay 20 bucks a month to use GPT and play with their web search capability ahead of others. So I think we should keep an open mind into what the new formats and environment might look like.
Speaker 1:Well, that 20 bucks is a lot more than just search, isn't it?
Speaker 4:It is, but that was one of the main reasons for a lot of people of just like you know what I want to go do the real time, and with perplexity. That's another big, big component of it. So I think you're absolutely right, there's multiple capabilities baked in, but you know search has been a part of it.
Speaker 1:So we only have a couple of minutes left. This has been great, awesome, super interesting. Let's go with a speed round. We like to sometimes end on a speed round and I think we're talking about. What I'm really interested in is monetization. Let's take it out two years, one to two years what do you think are going to? How are these AI we'll call them engines going to monetize? So it's a guess. Obviously You've said yourself, chris, you don't know, but we're going to start with you, and the speed round means you know. Give us a quick test case answer. You know what you think of. Where does monetization come from?
Speaker 4:Yeah, I think organizations are going to end up paying for traffic, similar to Google, but it's going to look different. I do think we're going to think of visitors more organically and companies will pay for referral traffic from these models. I think there will be ads. Ads are kind of expected. Perplexity is already doing it, so I think you'll have normal ad formats that we've grown to know and love in the current ecosystem, but, I think, more exploratory. I think we won't even consider the fact that, oh, these are coming from these places and I'm paying a microtransaction for the fact that they're directing it to me. I think that's a potential direction of the world, lori.
Speaker 2:I don't think I have a very good answer. Just to be really honest, I don't. I mean, I do think they obviously have to be monetized, unless they're going to be funded by the governments, which I don't think we want. That either, that either. So I would have to say you know, agree, agree with Chris, it's got to be monetized by, by the businesses that want the traffic to some degree. Don't know what that looks like, though.
Speaker 1:And I'm sure Christian does have an answer.
Speaker 3:Yeah, I'd say I'd say similar. The first thing I think of is honey is honey. The browser extension that took 2% and just sat on everybody's affiliate programs. I think everything ends up at 2%. I think the displayable ads are fine, but the AI agent's just going to ignore displayable, so those will stay for humans. But where it's the AI agent or the AI personal agent who's scraping out the ads to show me real results, it's two percent coming from the merchant. So now we'll give stripe their piece, visa their piece, and the ai agent, the ai, the llm, their piece, and we'll be at six percent instead of 25. And then one more thing the website gets intermediated, since I'm into this whole intermediation thing. So Google's gone.
Speaker 4:OTAs are gone. Websites are gone. Yeah yeah, I agree with that. I think websites look like repositories of data for agents to take action. So the simplest is a transaction fee, maybe a smaller fee to book a meeting, book a chat. So I agree with Christian, it's going to be an interesting internet down the road.
Speaker 1:Yeah.
Speaker 1:It sounds a little bit like meta search, just a very little bit. But when Kayak started, you know, it was all pay per click, right, and whether it was 10 cents, 20 cents, 30 cents pay per click, and then they went to conversion and getting a bigger commission for conversion, and I do think it's going to go back to pay per click, but a tiny amount. So pay per click but tiny amounts of pay-per-click, but only certain people will do that right. The bigger companies will pay per click. You know, maybe it's today's OTAs. Instead of giving their money to Google, they'll pay per click for the AI engines where the smaller companies won't. So, once again, they're going to rely on those OTAs who have paid per click and get that traffic to make the bookings.
Speaker 3:But it might be. What is a click right and a click's always been obvious to us, what it is, today still, but in the future, like what is a click, it's the agents chatting back and forth.
Speaker 1:Maybe there's never a click. Well, there's going to have to be a company that will measure clicks, right, is that?
Speaker 3:a startup in the making. What do you do? You measure clicks on the internet.
Speaker 4:There we go we'll discuss that next time.
Speaker 1:Bruce can get busy building yeah, I don't do tech startups, chris, okay any last words.
Speaker 4:No, thank you for having me, love the conversation and, uh, appreciate having the chance to be on the show.
Speaker 3:Thank you, guys yeah, best of luck. I think it's a great space. I think I was just at itb last week and this is what everyone's talking. Everyone's talking about this because it's such a it's everybody's search, it's everybody's custom acquisition. So, um, everybody needs to pay attention to this. So, yeah, best of luck with everything you're doing.
Speaker 4:Appreciate you.
Speaker 1:Thanks, chris, great to connect with you and I'm sure we will do more soon. All right, thanks everyone. Thanks guys, until next time.
