Inuvo, Inc. (INUV) Earnings Call Transcript & Summary

Good morning, and welcome, everyone, to a very special event. This launch is -- the year of the ChatGPT and Google Bard have clearly signaled that the age of artificial intelligence has begun. What you're about to observe at this webinar is an AI that can generate marketable audiences at will without having to look up information in a database. We called this AI the IntentKey, and it determines and actions the key concepts behind the intent of your consumers. My name is Rich Howe, and I am the leader of Inuvo, the developer of the artificial intelligence that powers the proprietary experience we are about to share with you. For our shareholders, we filed our quarterly financials and accompanying press release yesterday evening. You can access both of those documents at our website. Inuvo was early to the application of generative AI within advertising, having invested nearly $50 million to date. Over the last 6 years, we've also witnessed the performance success that can come from the adoption of our technology, and I would encourage you to review a sampling of those case studies at our website. The Audience Discovery Portal we launched this morning is a place where information about any audience can be instantly generated. Like ChatGPT and Google Bard, the IntentKey AI that powers the portal is distinguished from all other types of AI by its proprietary model of the language. The reason this kind of AI is so powerful is because humans understand, experience and interact with the world through language. At Inuvo, we have taken that one step further and made the AI's knowledge immediately actionable for advertising. This has never been done before. The IntentKey is itself representative of the collective knowledge of humankind, having been trained on tens of billions of individual pieces of information. It has, for example, already read and associated virtually everything that has been publicly written about your product, service or brand and has formed a conceptual understanding of those things relative to marketing. The AI continues to refresh its language model, reading millions of new pages every single day. We built this experience to showcase a sample of this AI's capabilities. We invented the technology for 2 reasons. First, we wanted to solve the privacy issues related to using a consumer's information for advertising. And second, we wanted to unshackle marketing from the limitations associated with using identity-based technologies limited by their dependence on structured datasets. We believe we've accomplished these goals and, in so doing, have forever changed the advertising paradigm from being about who someone is to why audiences are actually interested. The days of using a consumer's data for digital ad targeting are coming to an end. This fact is self-evident in the ongoing decline and lack of scale associated with using third- or first-party targetable consumer cookie identifiers. The AI we have developed does not depend on persistent and intrusive user tracking. Rather, this AI has been trained on a more extensive set of public data that allows it to generate knowledge that did not previously exist. This could not be a more perfect evolution of the information age that began in the mid-20th century, and Inuvo foresaw this many years ago when we first launched the IntentKey. Aversion to change is a very human quality. However, as the saying goes, fortune favors the bold. And those of you who adopt this advertising technology ahead of your competitors will be the ones that lead this age of AI. If you could peer inside our language model, what you would see is a giant web of tens of millions of interconnected concepts. These concepts are abstract ideas, emotions, people, places, things, thoughts and beliefs. They play a critical role in all aspects of cognition. The audience for any product, service or brand does not need to be identified by the identities and demographics of the people within that audience. This itself, I know, is an abstract notion, given marketing has been based on this paradigm for generations. But think about it. What is important about a product, service or brand are the reasons why consumers have an interest in these things. For the first time ever, an artificial intelligence designed by Inuvo can instantly identify and action these reasons without knowing or wanting to know who those consumers are. This kind of transformational language-based artificial intelligence can and should power not just media but creative, content, market research, product design. Before we begin, I'd like to point out that for this demonstration of the portal, we have only provided a subset of the data our AI can, in fact, generate. Our clients get a custom version of this portal designed and optimized specifically for their business needs. What I'd like to do now is turn the webinar over to our Head of Marketing, Katie Cooper, who will walk us through all the information this artificial intelligence is generating at the portal. Katie?

Thank you, Rich. And hello, everyone. Again, my name is Katie, and today is a very exciting day. It's my pleasure to take you on a tour of the Audience Discovery Portal. Before we begin, I want to bring to attention again that everything you're about to see here is generated by artificial intelligence based on a language model that does not require intrusive or persistent consumer tracking, identity or data. And yes, you will be hearing this friendly reminder again later on in the tutorial as this is a key differentiator for our technology and the audience discovery capabilities our technology offers. So let's dive in. First things first, there are 3 key sections in the portal. The first is a free-form search capability. The second is a section about top audience categories that we see on the open web. And the last section, which is also represented in the ticker at the top of the page, highlights concepts that are trending up or down. As Rich said, the audience concepts are the building blocks of our AI. The abstract ideas, underlying principles, thoughts and beliefs within our language model that play a vital role in cognition. One thing to keep in mind is the data that you see here is actually changing all the time because the artificial intelligence behind this portal is continuously monitoring the audiences for any product, service or brand on the Internet at all times and identifying the things that are the most engaging at any given moment. So if you come back here in another day or so, you might notice that the categories or concepts have changed. While entertainment is #1 right now, it may not be next week. Our AI recognizes those changes in attention, and those changes are reflected here. Also, almost everything in the portal is clickable, so you can explore as you like. So for example, if we scroll down and look at trending concepts. Let's see. We can click on Katy Perry. And the AI is going to generate detailed information on the audience, it associates with Katy Perry. In the context of this portal, you can think about Katy Perry, the singer, as a product and think about the information that's been generated here as the potential audience that's interested in the Katy Perry product. We'll come back and revisit this view here later. We can also interact with the ticker at the top of the page here in the same fashion. This section displays again the trending concepts from around the web. You can hover and pause to scroll and then click into an audience detail for that concept. So we've touched on functionality and how concepts are the building blocks to the AI, so let's go to the top search bar here and dig into an actual example. I'm going to type in the American workwear retailer, Carhartt. So this brand is generally known for blue-collar workwear, coveralls, hardware and clothing, et cetera. So let's take a look at what our AI knows about Carhartt. As you'll see within seconds, the AI generates its assessment of what the potential audience looks like for Carhartt. Again, generated without consumer data. And I warned you early, this was worth repeating. At the top, we see some general information on audience size and the growth trend for this audience. But I'd like to focus on the circle area here because this is really where the differentiation of our AI is relative to existing methods of audience creation. It's worth pausing a moment to see how marketing has been conventionally done for at least the last 2 or 3 decades. With this guy. In the modern way of doing advertising, all the technology and systems typically key off of an identity, the identity of this person or people that look like him while continuously trying to add information on top of this person's identity. And the things in the blue circles here, income, age, demographics, interest, et cetera, this is the way it has worked. You just keep adding more and more information, and then you try to find a bunch of other people who look like this person because you know the same things about those people. And then that group of lookalikes collectively becomes the audience and who you end up marketing to. Also note the size of the circles. The equal size is representative of the equal weight this information plays in the makeup of this audience despite the actual influence that they play. What we are trying to do here at Inuvo is change that identity-based paradigm. Specifically, we wanted to create a privacy-safe technology that was independent of individual identity. So we went through language. And we said, "Language could be the proxy for an identity," meaning there are certain concepts that are probably associated with Carhartt, for example. And if that's the case, why don't we focus on trying to figure out what those concepts are and then work on improving the AI's understanding of those concepts, extending, if you will, the knowledge of each concept so that we can effectively get to the same place in terms of its ability to find an audience but without identity and, more importantly, successfully improving upon that data. So with that as background, let's explore Carhartt some more. This section displays a cluster of concepts that the AI has already associated with Carhartt, and the varying circle sizes represent the strength of that concept as opposed to what we saw earlier with the equal sized blue circles. There are actually thousands of concepts that our AI can associate. But for the purposes of the portal, we are only disclosing a set of higher-level concepts. Our IntentKey clients, however, receive dynamic just-in-time concepts. This is achieved by the AI reevaluating the associations based on what's happening in the world related to Carhartt, all the time. So think about it this way. The AI has read every page of content ever written about Carhartt. So hey, Carhartt, give us a call. A few interesting things about the Carhartt concepts we noticed right away is Jessica Chastain and Matthew McConaughey. You may recognize these names as Hollywood actors, but somehow they're associated with Carhartt. These interconnections are the AI's way of saying, "Hey, there's a connection here between Europe, Dearborn, Michigan, Jessica Chastain, Matthew McConaughey." Moreover, if we click into Matthew McConaughey and the dedicated concepts for his audience, the Carhartt model is actually inheriting these levels of granular sub-concepts as part of its audience makeup as well. These concepts go beyond what the brand itself may even know and generates connections that were previously unknown and with great influence and great relevancy. So how do our concepts in the portal manifest themselves? For Jessica Chastain and Matthew McConaughey, it turns out, were wearing Carhartt jackets in the movie Interstellar. The AI figured that out, which, let's face it, that's really cool. So what's the importance of that as it relates to the advertising side of the equation? I would like to highlight 3 actionable opportunities that result from our ability to define this audience based on a collection of concepts rather than a collection of identities. Number one, content. Consider the likely hundreds of articles written about the movie over the years. The AI can instantly associate Interstellar and Carhatt and decide to buy a media spot because the audience that's looking at those pages about Interstellar is an audience that should also be interested in Carhatt, and the placement of that ad has nothing to do with the consumer that's in front of the screen and has nothing to do with their identity. And therein lies the privacy component of this technology and the power of it to understand all of these concepts rather than use data and demographics. Number two, creative alignment. The concepts play a very important role in the ad creative development by illuminating themes and messaging from the learnings to incorporate into the creatives themselves. If we explore another concept here, like A.P.C. A.P.C. is a French fashion brand who, as it turns out, has done multiple collaborations with Carhartt. The strength of the A.P.C. concept could strongly influence the traditional Carhartt aesthetic, which again is more construction workwear, to something more high fashion and trend. This could be reflected in the ad created to garner interest from the A.P.C. concept audience. And again, these concepts would be inherited into the Carhartt concept. And lastly, partnerships. The learnings can highlight an opportunity for a brand to work with a strong concept. In this case, a celebrity like Matthew McConaughey.

Katie, I wonder if I could just add a couple of words here because, clearly, this is where the greatest differentiation between the way a language model-based AI like the IntentKey is from the traditional methods, which have always been based around identity and people. But as I said in the opening statements, when you think about marketing, it's not really about the people. We've made it about the people, those of us that have been a part of designing the data that gets used for marketing. We designed all of that around people simply because that was the way the technology allowed it to be designed at the time. It's really never been about that. It's really been about what people are thinking about a brand like Carhartt in this example. That's what really matters, is like what is the actual thinking of my audience about my brand or my products or my services. The Interstellar example that Katie gave is a perfect example of that. I mean what is it saying? It's saying that a bunch of people watched Interstellar and thought Carhartt's clothing, in particular, the jackets, if you go look at the movie, were cool. That's a good thing for Carhartt. These people think that their stuff looks really cool. That's the kind of audience you want to put your products in front of simply because they're predisposed to have an attraction to the brand already. It's really a way more powerful way of looking at the audience building problem because it opens you up to the ability to extend audiences in ways that have never been possible before and, to a large degree, have just simply been limited by very structured and limited datasets. There really isn't a lot of information that you can add on to a person's identity, measures in the thousands, maybe. This particular type of technology opens you up to millions. As Katie pointed out, when you are a client of ours, you get all of the associations. And Carhartt, for example, I know there's an association with them with fire-retardant materials. So if you think about it, it is a workwear. It's for people who are out in the fields, and they would want materials that are fire resistant, fire retardant. And as it turns out, that is associated with Carhartt. Again, another perfect example of how our AI would be able to find that audience. You can use thousands of pages talking about fire-retardant and fire-resistant materials. None of them, by the way, have anything to do with Carhartt. There's no mention of Carhartt at all in any of those discussions. However, that's exactly where Carhartt might want to put the products they have that are resistant to fire. So it's really a way to extend your thinking about audiences and have a technology capable of reaching micro audiences you never dreamed you could. Katie, I'll turn it back over to you.

Thank you, Rich. Well, with all of that as background, let's explore the audience for Carhartt. Our technology goes way beyond the general makeup of a prospect audience by using these concepts, but it does have the ability to generate demographic information, information that may appear familiar, but I want to stress this difference. Concepts, not demographics, drive audience selection. The demographics shown here are inferred from a collection of these concepts. So let's look at this. What the AI is saying here, which I find kind of interesting, is the split between male and female is basically half and half. But note the star here. What the star is telling you is an engagement rank. For female, this is a score of 76, with 100 being the highest relative engagement. So this means that the audience of females is a strong candidate for Carhartt retail products as they are very engaged online with this brand and their products. This isn't to say that males aren't engaged. But since there is no star, it means that the engagement rank is below 50 and, therefore, less engaged than female. This same engagement scale applies to all the demographic categories in this view, and these engagement ranks create opportunity to evaluate your advertising plans like age, for example. You could look at the younger age groups and see here that they are very engaged online with Carhartt. And you could ask, "Does my marketing plan take that into consideration." Or perhaps, the less engaged age groups here are an opportunity to action an underserved audience segment, especially since, in this case, for Carhartt, a strongly associated concept like A.P.C. ranked very high with the older age segments. The wealth of knowledge ready to be explored through these layers and layers of concept data is truly remarkable. So lastly, backing out to the main page. We also have a section for industry audience trends and as well as an FAQ. And if you have a question not shown here, use the multiple contact buttons to reach out and learn more. We've also included some helpful resources in the portal menu. I recorded tutorial similar to today's webinar, case studies highlighting our ability to discover and action audiences others can't and, again, our contact. Please don't hesitate to reach out to us for your customized insights about your own product or brand. And with that, I thank you for taking the time to learn more about the Audience Discovery Portal. This concludes our tutorial, and I'll be turning things over to Natalya Rudman for Q&A. Happy discovering.

Thank you so much, Katie. As a reminder, if you have a question about Inuvo's Audience Discovery portal or Inuvo's AI, please submit your questions to the Q&A box. Let's get started. Our first question is, how is it possible that you could have AI as powerful as ChatGPT or Bard?

Yes, that's a great question. I guess the answer to the question, Natalya and whoever asked it, was we've been doing AI for many years already at Inuvo. And I think as I may have indicated in my preamble, our founders and our Board members and many of our executives are from notable companies who built most of the methods that are currently used to do marketing and advertising and the data businesses that support that. So we had maybe a jump on others in terms of our thinking about how to solve the privacy problems and how to leverage the latest and greatest technologies associated with doing that. And it was clear to us even many, many years ago when we first started developing this technology that a language model-based generative artificial intelligence technology would be the way to do that. So we didn't just wake up and think that, that was the case. It is our experience that led us to that. And I think history is proving us correct in this regard because we do see these announcements coming out now from ChatGPT and Google Bard. However, as I've mentioned before, we are the only company that I'm aware of that has applied this to the marketing use case.

Thank you for that, Rich. Our second question, how long does it take to get an ad campaign going?

Yes, that's another great question and another very huge advantage of technology like this. So in traditional advertising, we all know the construct. We take a look at all of our clients. We go buy a bunch of data about them, consumer data. We overlay it, we segment it, we do some analytics on it, we try to figure out where we think we need to go target and then we go purchase third-party datasets that are aligned with those identities. There's a lot of work, a lot of money that goes into doing that. None of that's required with this technology. This technology effectively eliminates the need to do any of that. It automatically figures out what the audience is. You saw that today. You can see it at the portal. And then it automatically figures out where the media should be placed based on its understanding of the product, service or brand. So the answer to the question is in most of our client the implementations is a few days, it really just takes the AI to be focused on some objectives, which we work with our clients on. So within days, the AI is knowledgeable about where the best audiences are located, and we can begin running campaigns for clients.

We have another question that came from our audience. As a potential customer, what services do you offer to complement this AI?

We offer a full set of services. So we can run the campaigns and use our AI to be driving that campaign strategy. If -- for our agency clients, we can plug our artificial intelligence into whatever campaign systems that they are actually using so that we can empower them to be successful with their clients.

Another question from the audience. What is the difference between what you are doing to gather information about consumer behavior different from others in this space?

Well, the first difference is we don't use identity. So as I mentioned earlier, everything about marketing for at least 2 generations has centered around the person. If you could look inside the databases that are used, you'd have a picture -- not a picture, but you'd have a number that's associated with, Rich Howe and then there'd be this long list of attributes that are -- that define me, my age, my income, my family, anything else, my DMV records, anything that can be found gets added to that. And then that becomes the targeting record that gets used to do advertising. I'm sorry that I went off Natalya. What was the question again, just because I went off explaining how it's actually done now versus what the question was. Remind me.

Yes, no worries. Yes. The question was, what is the difference between what you're doing to gather information about consumer behavior, differently from others -- yes, how are you doing this?

So that's the way it's done now. Our method is different, and it's very much like ChatGPT and Google Bard. The technology that we've developed has read tens of billions of pages of content on the Internet, and it has interpreted that content and understood the relationships between pieces of that content, the words on the pages, the pages themselves relative to other pages. And the consequence of that is a large -- if you could look inside, a large model of the language at the conceptual level where there's connections, if you will, between all of the concepts. There's actually 25 or so million of them in our language model. Not that, that matters, but more than enough. So there's really an unlimited way for the AI to kind of find its path towards these audiences. It takes a lot of computing power and processing power and a lot of really smart AI-based scientists and engineers to be able to build something like this, let alone actually be able to implement it. But that's the fundamental difference, is that there's really not consumer identity-based data involved in anything we do. It's all public information.

Next question. Can you apply this technology to market research and product development? And is this something that you offer?

Yes. The answer is yes, and I'm not aware of any other technology that has the ability to do that. So the answer is yes. If you think about it, we've been constrained to some degree, again, with the way the environment was created. Where if you're launching a new product, you're at a somewhat disadvantage in terms of trying to understand what does the audience actually look like for that. And maybe more importantly, how should I design this product, a product that doesn't -- hasn't existed yet so that I'm aligning the features of that product with what ultimately the audience is interested in. Many of us would probably do surveying to do that or at least to get some perspective on that. That's not necessary with the IntentKey and the AI we've developed. We can point the AI at any product, even products that are conceptual, and it will help refine and define not just the audience for that, but what the audience is interested in related to that product, meaning the actual features of it, which should be part of the design. So it's actually a really great added benefit, and I'm glad that question was asked because it's not just a media technology. Of course, it was designed to be a media technology. But there's so many other areas: creative ideation and content creation and this, product research and design. There's so many other capabilities that the tech can facilitate.

Moving ahead. Is the AI able to place an ad online?

Yes, that's effectively what it does. If -- go back to the Interstellar example that Katie said. Not only would it have figured out that there's an audience of possible purchasers of Carhartt's gear who were watching the Interstellar movie, but it will know all of the pages on the Internet. There's probably hundreds of them for that one, where Interstellar is being talked about that become good places for a Carhartt ad because, ultimately, the people in front of that screen are interested in Interstellar, which means they're probably interested in Carhartt gear because they watched the movie, and it was in there and they thought it was cool.

Absolutely. Next question. Can your tech also generate hyper-personalized creative in real-time based on targeting?

It doesn't -- the tech has not yet been designed to generate creative, but the concepts that are the result of the AI's definition of the audience for any product, service or brand, absolutely should be used by the creative agencies that support the companies we work with to help them with ideation related to the creative. And I would submit that it opens them up to a whole different way of thinking based on a whole new subset of information. A lot of creative ideation is just that. It's a lot of people think -- sitting in a room, trying to come up with some great ideas. The concepts that we associate with the with the brand, product or service are, in fact, the stimulus for that ideation. So it's a very powerful component of the creative. We don't do creative at Inuvo. We partner with companies that are good at that.