Loading summary
Transcript53 lines
- [00:00]
A
Foreign. Welcome to Ad Exchanger Talks, the podcast devoted to examining the issues and trends in advertising and marketing technology that matter most to you.
- [00:32]
B
I'm Allison Schiff, and thanks for listening to Ad Exchanger Talks, your friendly neighborhood ad tech podcast. My guest this week is Deborah Ajo Williamson, founder of Sonata Insights and its chief analyst. Sonata is a consultancy that helps marketers and media companies understand how AI is reshaping consumer behavior, media consumption, and advertising strategy. We'll get into all the good and not so good stuff, including how AI is changing creativity, why she thinks we're entering the age of AI media, the evolving rules and expectations around AI transparency, how brands should organize a budget and experiment to get truly AI ready. And that's just for starters. If you want more, and I know you will, Deborah is keynoting our programmatic AI event coming up May 18th through the 20th in Las Vegas. She'll be talking about what AI looks like from the consumer perspective and what that shift could mean for targeting platform, platform power, and the evolving role of programmatic advertising. So snag your ticket. In addition to Debbie, we've got a great speaker lineup that includes folks from JPMorgan Chase, Microsoft, AI People Inc. Horizon Media, WPP Media, Good Apple, and much more. Podcast listeners get 10% off the price of their ticket when they use the code POD10. So see you there. Hey, Deborah, welcome to the podcast.
- [01:51]
A
Hey, Alison, it's great to be here.
- [01:54]
B
What's one thing about you that not a lot of other people already know and that you couldn't easily find out just by looking at your LinkedIn profile?
- [02:02]
A
Ooh, that last part is a tough one. But I'm going to go with it anyway, because you have to go all the way to the bottom of my LinkedIn profile to read about it. I live in Seattle and I love recommending restaurants here. Um, I am like your favorite person if you want to know where to go to eat. And there's this little hidden place in the Pike Place Market that everybody goes to, but not. Well, they go to Pike Place Market. They don't go to this restaurant. And so I always recommend it. It's called Matt's in the Market, and it's upstairs in a. In a building adjacent to the market. And so that is maybe not something so interesting about me, but interesting about the Pike Place Market because it's a huge tourist destination because. And Matt's in the Market is one of the best restaurants in the area. So that would be my recommendation. If you come to Seattle, you're a
- [02:53]
B
good Friend to have. I'm the kind of person that has a few friends. I have a few friends who are foodies. And so I just follow them around New York City. And so I know a lot of good restaurants, but I don't know them because of me. I know them because I'm riding the coattails of others.
- [03:08]
A
That works.
- [03:10]
B
So I know that the first thing I asked you was to share something about yourself that wasn't on your LinkedIn. You did, but that's totally fine. But I was, of course, also looking at your LinkedIn to get ready for our chat today, and I noticed that you're a judge for the Seattle AI Film Festival, which is a new thing. It's organized and Produced by this B2B podcast and newsletter called Culture and Code. And they focus on generative AI for creative professionals. And what they say about themselves is that they're helping storytellers copy collaborate better with AI and with each other. So there's a lot, a lot of tension right now in creative circles about AI, questions about training, data and credit where credit is due, compensation. Like, what counts as real creativity if a model's involved. So I know that the festival's coming up. You have the dates. I believe there was one last year. So from your involvement so far, what have you seen that gives you options, optimism about AI and creativity? And is there anything that kind of gives you pause? And, like, could you tell that the films were using AI? Yeah.
- [04:24]
A
Oh, gosh, what a wonderful way to start this podcast. So the Seattle AI Film Festival was founded by my former coworker at eMarketer. So I was an analyst at eMarketer before doing what I do now. But John Gaunt worked with me as another fellow analyst covering mobile, and he has moved on to AI, as I have, and he focuses very much on the creative side. So he helped to launch the Seattle AI Film Festival. So I enjoy so much the opportunity to work with John again. So the film festival had its first run last year at the Seattle center in March. It's coming again in 2026 at the end of August. And what really blew me away last year was how incredibly artistic these films were. I had a vision in my mind that I was going to be seeing a lot of kind of, you know, maybe clunky things or maybe, you know, five, 15 fingers, because video wasn't very good back then. But what I wasn't really. And I did see a little bit of that, but what I wasn't prepared for in the winners of last year's film festival, was the absolute level of artistry that took place. And so what that tells me and what, where I think about it with regards to the advertising community is that there's so much focus on using AI to become more efficient, to save money, to increase your productivity, to increase your output. I think those are all okay reasons to use AI, but the most important reason to use AI in any sort of creative pursuit is to do something that can't be done or to express creativity in a different way than you could in other methods. And that's what I think this film festival shows. And what I'm expecting to see this year is the beauty of using AI, of the ability to prompt, to create imagery, to create visuals, to create something that really touches you. If advertisers remember that, I think that's going to make the AI in the creative process much more natural fit and much more accepted by consumers. Well, I don't know, but I'm really excited about the film festival coming up.
- [06:56]
B
Do you think you would have been able to tell that AI was being used? I guess in cases of many, many fingers, you would know. But if you didn't already know, if you didn't know you were judging an AI film festival and you just saw these films, would you have known that they were AI generated or largely AI generated?
- [07:14]
A
Not necessarily. Some of them look like animated films. Like there was one that I remember that was. It looked like stop motion animation. Each scene was captured, but it was done in prompting. Others were very photorealistic and looked like a film trailer, like you would see in a theater. And what I think is really going to be incredible is that what we saw last March was very early in the use of AI to create video. And we're now a full year ahead of that. Plus by the time we get to August of this year, the models have gotten so much better that I think that the creativity, the realism, but also the ability to just generate worlds or thoughts or emotions, things that you wouldn't expect from AI. I'm really excited to see what will come in the next generation of films because I just know that the filmmakers have gotten better and I know that the processes and the underlying techniques have also gotten better.
- [08:26]
B
Well, as things get better, you know, it's exciting to think about all the cool things you can do, but then you also have to pump the brakes a little bit because there's the policy side governance, and so there's this parallel track happening right now. Of course, you know, it's about labeling things correctly on the creative side and Giving credit where, you know, directors deserve it or whatever. But you have the IAB and other groups trying to hammer out some basics around AI governance for advertising, including, like, how and when AI generated or AI optimized ads even should be labeled. And so my question is, like, as all, as all of that is still being worked out, what do you think is the right level of transparency for people seeing ads that include AI? Like, do we need big in your face labels or just like a little symbol or something like a quiet disclosure in the background? Or does it have to be somewhere in between because you don't want to overload people? Then again, you don't want to hide the fact that you're using it.
- [09:30]
A
Exactly. And it is a big debate in the industry right now. And since we're talking about disclosure, full disclosure, I worked on research that underpinned some of what IAB came out with in terms of its guidance for disclosure of AI use in advertising. Just to quickly recap the research, we surveyed young consumers, Gen Z and millennial consumers, and we compared that to advertising executives. And we asked. Essentially what we found is that younger consumers are much less positive about the idea of a generated advertising than advertising executives thought they were, and that the gap in opinion had actually grown since last year when we did the study a year ago. And so we also found, of course, that more people are exposed to AI, so they're more familiar with it, and more ad executives were using AI. And so what that led us to conclude is that disclosure actually can be beneficial. We did find in the research that people did feel a little bit more positive towards the advertising, but more importantly, they were more likely to make purchase, or they were either more likely to make a purchase or it had no impact on their purchase likelihood. And that, I think was really compelling for what IAB is coming out with. That said, I mean, you raise a really good point, which is that advertising that is generated with AI, I think eventually AI is going to be part of all of advertising generation. So just like we think about CGI or we think about Photoshop, we don't necessarily label that something was created with Photoshop or that we used CGI in the ad creation process. And so there's this question going, well, eventually, if people are going to always use AI, why should we have to label it? What it comes down to for me right now is consumers don't want to feel misled. They don't want to feel like they were tricked. And I think that can go both ways. So you're seeing some advertisers actually take a stand and say, we are not using AI. We're going to label our ads as being not created with AI, which I think is good. And then I also think, though, that advertisers that are using AI in a way that potentially could mislead people. A great example would be an AI generated spokesperson who's giving a testimonial for a product and saying how much they love that product when it actually isn't a human. That should probably be disclosed. So that's where I think it gets interesting. There's not a way to do a blanket disclosure, but it really kind of comes down to is what you're doing potentially going to create the wrong impression or mislead consumers? And I just always feel like advertisers should be on the side of the consumer. And I hope, you know, I know that's not always the case with advertisers, but I make hope is that when they use AI that that's what they will. That's the goal that they'll be looking for and they'll be shooting for.
- [12:36]
B
It absolutely is about trust. I remember reading something a little while back about people that were communicating with a therapy bot, and they, well, they didn't realize it was a therapy bot. They thought that they were communicating with an actual therapist. And when they found out that it was a bot, they felt terribly betrayed. There was some research that was done to try and ascertain, like, whether people would engage in therapy if they knew that it was a bot beforehand and would be more okay with it if there had been some disclosures. And the answer was yes, Actually, there are some people who are never going to do it, but there are others that are actually fine with it. And they don't mind that interaction so long as they're told from the beginning and it's not a, you know, unpleasant discovery once they've already opened themselves up to a chatbot.
- [13:30]
A
I mean, yeah, it makes complete sense. And people don't want to be misled. And if people know what they're getting into, recognize what the parameters are, I mean, why wouldn't you want to support that? You know, whether you're an advertiser or you want me to market a therapy chatbot, or whether you're, you know, a chatgpt, you know, keeping people informed and is. It seems like basic table stakes for me when it comes to marketing.
- [14:01]
B
You brought up, though, I want to go back to something you said, which is really interesting. This trend right now of brands explicitly highlighting that they don't use AI in their ads. And that is a way that they're trying to build trust and show that they're actually really creative and they care about humans and all of that. Like, there was. There was that pledge that Dove made, I think it was in 2024, like never to use AI to create or alter images of women in advertising. That makes perfect sense because they already had that campaign, the Real Beauty campaign. So they just changed. Not didn't change the name, but they added a layer to it, like keep beauty real.
- [14:38]
A
Yeah.
- [14:39]
B
Aerie by American Eagle. They made a pledge like that. Lego Polaroid, I think Heineken, like many others. And then you have Coca Cola out there unapologetically launching these fully AI generated holiday ads two years in a row. So I don't know, I kind of feel like we're at this transitional point where it makes sense to. For some brands to cry from the rooftop that they're not using AI. But by the same token, like you just said, AI is eventually and maybe kind of already is, unless you're taking a particular stance woven into advertising. So, yeah, I mean, do you see that as a permanent stance that brands can take, that they're just going to be out there saying, we don't use AI or is it kind of like a holding pattern that they're. And that they're in. Until brands get more comfortable using AI in a way that feels authentic, and then people are also more comfortable.
- [15:38]
A
I mean, I think eventually, like I mentioned, AI will be part of the creative process in some way for any sort of advertising creation. So we are kind of in this middle period where there's a general consumer. Some consumers feel very skeptical toward it, and maybe others are more accepting. And the same with advertisers. Some are very skeptical and believe that AI should not be used in advertising, and then some are leaning into it. You use some really great examples. I think that really, again, it comes down to me to remember that if you are using AI, use it in the interest of creating something new and different and hasn't been seen before. I think one of the biggest missed opportunities that Coke had was it's been using AI to recreate what is a beloved holiday classic ad that that has been around for years. And I look at those and I go, well, why didn't Coke just do something you haven't seen before that's different, that leans into AI or maybe demonstrates that, you know, AI in a certain way, that that relates to the brand of Coke, rather than, let's just use it to, you know, make the polar bears different, you know, you know, interact in a certain way and, you know, make the train go a certain way. I don't know, it. It feels to me was just a missed opportunity. And then with the brands that are in beauty and fashion, I mean, it makes sense that these kinds of companies would want to take a stand or maybe lean away from AI because they're marketing products for people, things to put on your face or your body to make it feel about things to put on your face or body clothing. So you see a lot of fashion companies using, saying that they're not using AI generated models, for example, because they want to show real human beings wearing their clothing. That makes a lot of sense to me. So it's an interesting middle ground.
- [17:48]
B
I mean, if you're going to advertise lip gloss or a beauty product, you want to advertise it to someone who has lips or cheeks or skin to rub the lotion into. By the way, I realize I'm remiss because I immediately started asking you questions and then I skipped over the part where I properly introduce you. You were a longtime analyst and a researcher at eMarketer, you mentioned that. 12 years. And you founded Sonata Insights last year. And so now you're advising businesses on how emerging technologies are reshaping, transforming consumer behavior, media consumption, media strategy. You also have a substack called the AI Ad Economy about how AI is changing advertising. But what made you decide that this was the right moment to go out on your own with Sonata? And what are the main questions and concerns that you're helping marketers and media companies work through right now? And what are they asking you?
- [18:48]
A
Yeah, yeah, really awesome questions. So I spent my entire career at the earliest stages of major digital transformation, all the way back to when I was a journalist like you at Ad Age, covering the first ads on the Internet. And, and then at eMarketer, covering and realizing that social media was a profound change in the way that consumers behave and the consumers interact with each other, and that was going to change the way that marketing and advertising also was done. So I was one of the first analysts to pinpoint that social media was going to be such a big phenomenon that it turned out to be. But after as many years as I spent doing that at eMarketer, I. I realized that something new was on the horizon, which we've already been talking about AI And I realized that not only was AI going to change the internal processes of companies, so like, we've just been talking about the way that you generate advertising, the way that you buy media, the way that you analyze the results of your advertising, your marketing initiatives, using AI to make that more efficient, that's important. But for me, I always go back to the consumer and I pay attention to consumer behavior. And I noticed and I realized when I launched Sonata in 2024 that consumers were changing the way they were behaving again. And using platforms like ChatGPT, like AI overviews, like Copilot, to search and get information and make purchase decisions in new and different ways, and that these platforms were also going to have an impact on marketing and paid advertising. So my company is all about helping businesses to understand and make the most of this dramatic change that is happening again in consumer behavior, in marketing and in advertising with the lens of pay attention to what's happening with the consumer audience around you and make sure that your marketing techniques, your marketing tactics are up to speed with how consumers are changing. And when you ask about questions that I'm getting asked, there are a lot. And that's what I try to do every week in the AI ad economy, which is all about understanding where consumer behavior meets advertising in AI environments. So there are a lot of questions right now about how do I show up organically on platforms like ChatGPT. You get into this idea of AEO and Geo and how does that relate to my search spending and my search marketing and my search advertising? I get a lot of questions about the paid advertising model that's developing really, really quickly in again, ChatGPT being the most prominent example, but other platforms as well, where you can actually buy advertising that meets people when they're starting to make decisions in these platforms, like where does the money come from? Do I stop search advertising and put my money into ChatGPT? Where does this impact social media? How does retail media fit into this? So there are so many questions right now about how this new form of media, which I'm calling AI media, is transforming marketing and then also how it's transforming consumer behavior. So lots of things to think about, lots of things to research and analyze. And it's changing by the day, by the minute, it seems. And that's an exciting time for me. I love that. I love that early days of social when we didn't know who the big players were, we didn't know what the ad formats were going to be, we didn't know what consumer behavior was going to look like, but we knew it was going to be something big. And I think that's where we're at right now. With AI and AI media.
- [22:49]
B
So AI media, that's a term that you used in a column that you wrote for us. AI media is already here. Here's what marketers need to know. So we're going to take a quick break, but when we're back, we're going to dive into that and I'm going to ask you some of the questions that your clients are asking you so that our listeners can benefit. So stick with. All right, we're back. And before the break, I teased a column that you wrote. AI media is already here. Here's what marketers need to know. So in that piece you were describing ChatGPT and Copilot and Google AI overviews, Amazon's Rufus, that whole crowd, as a kind of new quote, unquote, AI media form of new media. So basically AI platforms that are starting to look a lot like media companies because they're selling ad space. But it is a lot more than that. That's kind of reductive. So unpack your thesis a little bit and explain what makes these AI assistants and AI native interfaces like media companies and not just like tools.
- [24:08]
A
Right? Yeah. So I mean, this is a really profound difference and I think would be helpful to just start with a definition of AI media. So the definition that I work with when I use the term AI media is that these are ad supported experiences inside of AI assistants, AI enhanced search and AI native interfaces where people ask questions, research and make decisions. That definition could evolve as this concept of AI media grows and changes. But that gives me us a lens of what to think about. First of all, the biggest thing I would say is that AI media is more than chatbots. It's not just ChatGPT or AI overviews or Gemini or Copilot. AI media is also ads that are contained within Rufus, the AI portion of Amazon, ads that are contained within Sparky, which is the AI portion of Walmart's web mobile app. There is a whole long tail of AI apps and services that also take advertising. And there are companies that are springing up to allow businesses, kind of like in an ad sense or an ad network sense, buy ads in this long tail of smaller AI apps. So this is a big, a potentially large business that is developing kind of before our eyes. And it is to your point, for advertisers, something to recognize that these are not just tools. They are tools for many people. You know, people use them for coding, they use them for, you know, constructing spreadsheets for creating graphics. You know, we talked about AI for generating advertising. I mean, These are all wonderful tools, but they are also media surfaces where people, consumers are going and they are getting information, they are doing research, they are making purchase decisions based on information and what they're getting back. They're making medical decisions, they're making financial decisions. Lots and lots of consumer activity that has been typically taken place and mediated in other environments is shifting towards these platforms. And so that's what makes them media, and that's what makes them a really interesting, compelling place for advertisers to start to think about interacting and engaging.
- [26:47]
B
Well, because they're, they're media, but they're not just another form of media. They're kind of their own beast. So you made this interesting point in your piece that AI doesn't just capture intent, it also, like, kind of reshapes it and gives it momentum, accelerates it. And that's because people can do all the things at once. They can comparison, chop and look at reviews and then narrow down their choices all in this one interaction. Instead of, we've all had that experience of a million tabs and you don't know which one you were searching for the flight information in. And it's a whole thing. So it's really a new way to discover information. It's also not just another form of search. It's its own thing. So is there anything that marketers need to unlearn from the classic search mindset? Because it's not just copy and paste in terms of strategy, so anything they need to basically forget, because if they do it in this way, it would set them back in, I don't know, whatever you want to call it. The agentic era.
- [27:50]
A
Yeah. Well, the first thing I would say
- [27:53]
B
is
- [27:56]
A
think about it as looking for where the intent is forming, as opposed to thinking about this as a place where you just, you know, build your keyword list and you place your ads against a list of words that you want your ads to appear against. It's, it's more, it's definitely much more nuanced than that. It's about figuring out what sorts of prompts, what kind of language are people using to describe what they're looking for, making sure that the content, your organic content, matches those, those, those prompts, but also making sure that your advertising and the advertising thing, the terms that you're using just to place your advertising, match that as well. So I would say, first of all, yes, unlearn the idea that this is keyword based. This is much more nuanced than that. The second thing I would say is that this is the way that right now I'm seeing advertising and media work. The best in AI platforms is when you're trying to reach people at the middle stages of the funnel. So not so much upper funnel, although there are opportunities for introducing new products to people, which happens all the time. But when people are getting into that consideration and maybe conversion decision phase, these are places where consumers are showing that they are very open to receiving advertising. So that's different from search. Right. Search is at the very bottom of the funnel, getting people to make a decision to convert. Right. In some way to buy that airline ticket to go, you know, go and, you know, go to Target and, and make that purchase. You know, the plant that you want to put in your outside of your house. Sorry. Thinking literally about things that I was thinking about this weekend. But, but, but the. But if, yeah, if I'm doing research like, oh, this is the front of my house and I want to put new plants in chatgpt, help me figure out which types of plants might look good in the front of my house. Those are opportunities where people are just starting to think and consider and make decisions. Really strong opportunities for advertising. And that moves people kind of at a different place in the funnel than search advertisers are typically used to.
- [30:29]
B
It also requires a really, a new way of thinking about advertising for another reason. Because if you have AI as an agentic layer that's helping people make decisions, brands don't just need to show up there. They also have to appeal to the model because it's the model that's helping make and shape the recommendation that will change the consumer's behavior. So in a way, yes, you're advertising to people in a new way in that environment, but you're also advertising to AIs because you want to be the information that they eat in order to make the recommendation. So the person buys your plant or planter or whatever.
- [31:11]
A
Yeah, I mean, it is interesting. So it's a little bit challenging to unpack because what companies do on the organic side of AI, so AI visibility, aeo, that is very important. That is what we're talking about, which is how do I show up organically, how do I influence the model so that it will hopefully show my brand, my product or whatever as part of the answer. You can't buy that influence. You can't buy advertising to influence the model like that. You can do smart things to optimize your content on your website, your product descriptions, your schema, your earned media, all sorts of things, be present on Reddit, where These bots are known to go and get information. Those are all good organic level things that brands can do. But the paid advertising that they place within a ChatGPT is not designed to influence the model. And ChatGPT, for example, is very, very distinct about how whatever ad might appear at the bottom of your search result of your results or your prompt interaction is not triggered or related to what's up here in the organic part of it. And that's a really important distinction. I think advertisers need to know you still need to be there. But the ads that you buy on these platforms, don't tell the platform, don't tell ChatGPT or whatever. Put my brand in the organic response, because they're not going to allow that. Oh, for sure.
- [32:52]
B
You just need to be doing both things as a brand and it's just good brand health anyway. You should be posting blogs about your cool products and getting real reviews from people with lips and all of that stuff. Absolutely.
- [33:07]
A
Just like in search and social, organic is an important component of your marketing in those types of media. And so it makes sense that organic is important in AI media as well.
- [33:20]
B
What do you think, by the way of OpenAI's projection for itself, that they're going to be generating like just over $100 billion in ad revenue by 2030? Because it's. To me, it does seem plausible because Meta makes somewhere between 40 and 50 billion a quarter in ad revenue, depending on the season. Google makes around 70 or 80 billion a quarter. Amazon brings in like high teens to low 20s a quarter. So 100 billion a year for OpenAI within four years does sound like an achievable goal, even though it sounds like a wild headline to read it right now.
- [33:56]
A
Yeah, it is. And I think most people would say that it is an unachievable goal. I would say that there's a lot that OpenAI needs to do in order to get to that point. And I think the first thing, though, we need to acknowledge is that this is AI. Everything is moving much, much more quickly in AI and AI media. Then it moved in search and it moved in social. So it took, you know, it took years for Facebook revenue to really start to ramp up. Long after brands created organic presences there. Facebook didn't, you know, it didn't come out of the gate immediately, being the huge revenue phenomenon that it is, I think that OpenAI will move more quickly. It's using AI to create its ad products, just like Meta has used AI to improve its own ad products. So all of these things that OpenAI is doing to launch and bring advertising into this platform are making the timelines shorter. Do I think it's that much shorter that in four years it's going to become one of the most significant media platforms, digital media platforms there is what needs to happen. So, first of all, OpenAI needs to have a lot more users. ChatGPT used to have a lot more users. People have established that it's, it's growing, but not as fast as it should
- [35:33]
B
be or could be.
- [35:35]
A
There's an interesting dynamic where that chatgpt has both a paid tier where you pay for a subscription and don't see ads, and a free tier. So it has a really interesting balance. It's going to need to strike where it wants to get people to use the platform enough that they see advertising, but maybe not so much as they pay the subscription fee to avoid the advertising. So it's going to have to sort of balance that. You know, it's how much, you know, where is the little lever there or the dividing line that happens where someone is so enamored with it that they're willing to pay the subscription fee and then the advertising doesn't even appear for them. And then I think another thing needs to happen is on the advertiser side, advertisers need to get up to speed really, really quickly. Again, that's what I'm all about with my newsletter, is helping people figure out how to get up to speed. But advertisers don't necessarily move as fast as tech companies want them to move. They don't shift budgets that quickly. They plan a year in advance. They can't necessarily go and pull millions of dollars from search advertising and plug it over to ChatGPT. It involves a lot more than just someone saying, oh yeah, let's just go do that. So I think the advertising world has a lot to catch up on as well. It's going to be really interesting times. But at the end of the day, I think what that hundred billion dollar figure tells me is that what OpenAI is doing right now is not a test. This is the launch of a major advertising business and they have significant goals in this and significant and they have significant resources to potentially be able to achieve those goals. So I'm all in it. I'm so excited to see where this goes and what happens next.
- [37:28]
B
Well, what specifically does a marketer need to do in terms of organizational readiness to really embrace AI and AI media in a way that, you know, really works and isn't just, I don't know some. Something they're experimenting with. Like when, when you look at brands and agencies that seem the most AI ready to you, what. What kinds of changes are they making to their org or their leadership? Like what actually looks different inside of those companies and what are they willing to do to get themselves ready? Even if it's kind of uncomfortable along the way? Because everyone hates change. It's not y especially organizational change.
- [38:09]
A
Yeah. I mean, so there's a few things. So first of all, I think having clear ownership of your AI initiatives within your company, having a leader or a team or an organizational oversight of your AI initiatives really makes sense. We talk about the proverbial silos that happens in AI as well, where you know, someone's doing something AI over here and I hear and here and here. But having someone who can oversee all of that and understand where the budgets are going and then have a lens also into where your consumer's behavior is going. So that would be the first thing. Have someone who can have clear ownership of AI initiatives. This is also moving really quickly. So a willingness to experiment. The window for experimentation is actually really going to be short. So companies really need to start now and figure out where does do some pilot tests, set aside a little bit of budget to place ads. If not, if you can't get into ChatGPT, look at something else. Advertise on Copilot, try Rufus, try Sparky, try that long tail of AI apps, Get some learning. But so have that appetite for. I'm. I'm open to seeing what can happen if I try advertising in some of these platforms. And then from there, I mean, being able to make decisions quickly. Again, things are moving so quickly that maybe you're not going to have all the data that you need. I remember this from the early days of social. And even in the later days of social, people always complain about measurement. Oh, we don't have the right enough information. Yeah, we don't have the right information. Except that we don't have enough information. We can't buy ads on the way we want to buy. It's not a cost per click, it's a cpm, whatever it is, get used to that. Understand that the learning that takes place is more important than getting all of the data and every single piece of information that you think you need. That I think would be kind of my third takeaway for marketers to think about. Don't worry so much about the data. Focus on the learnings.
- [40:32]
B
Yeah, just mess around. Roll up your sleeves. Where is the budget going to come from though, like when marketers are thinking about bringing AI media into their strategy. And maybe it's just a pilot test for now, but what bucket is that coming out of?
- [40:47]
A
Good question. We asked similar questions back in the early days of social and I remember people saying, oh well, the TV budget people will take some money from their TV budget and put it into social. People said search budgets would go into social. What ultimately happened was social budgets went into social companies. Literally. They did a little bit of experimenting and then they realized, oh, we actually need to have budgets specifically for social. But where we're at I think right now with AI media is that most advertisers think about this as being aligned with search and they often call it AI search. We've talked about how it's not exactly like search, but it as a lens for how you think about the advertising spending. It's probably where I think budgets will come from in the early stages. Do I think that's ultimately where what will happen? No, I don't think, I mean, I think search will still continue to exist as AI media grows. So we'll be interested to see how much companies shift over. But as a starting point, I would think search budgets make the most sense. Where to begin your initiatives and then from there there could be ways to integrate social. We haven't really talked about the social aspects of this and how they could grow, but I do believe that there could be more social ways to engage with ChatGPT or other platforms. Right now it's kind of like a one on one. You know, you're you and the thing and the chatbot talking, but sharing chats, interacting with people, sharing videos that you make with, with AI. I mean these are all social experiments and social activities. And you could see an overlapping sometimes someday with social media, not yet, but maybe in the future.
- [42:39]
B
The concept of buckets is kind of interesting because I'm thinking of how I as a tech journalist covered mobile. When I first started at Ad Exchanger, I was the mobile reporter and I actually funnily enough was also the cross device reporter. Like that was actually my beat. It's so specific. But 12 years ago it was kind of new and mobile was also still really getting its legs. And so it made sense to have me be, you know, this reporter of a very specific, on a very specific topic. And now you would never have a mobile reporter. Mobile is diffused into everything. And so I, I do think AI is very much like that because, well, we have an AI reporter now. Joanna Gerber on our team writes a lot about AI startups. It makes sense to have someone focused on that, but. But it makes sense for all of us to know about it because it's, it's everywhere and it's part of everything. So at a certain point, I do wonder whether you might not really have search money going into it. It's just retail media dollars going into it and video dollars, and it'll just be, you know, there in the ether.
- [43:50]
A
Yeah, absolutely. And the same thing happened with social. When I was an analyst at eMarketer. I mean, at the beginning I was the social analyst. I wrote all the research about anything having to do with social. But within a few years, every analyst was writing about social as it related to their coverage area. Social and mobile, social and search, social and financial services, social and retail. So you're right. I believe AI is just such a foundational part of, of marketing from within companies, but it's also very becoming, very foundational to how consumers are getting information and making decisions, and that impacts all forms of media. So you make a really, really good point about that.
- [44:39]
B
So we're nearly out of time. So I had one last question. I wanted to end on a bigger picture, so to zoom out. It feels like AI is being talked about in two totally different ways right now, but at once. So on the one hand, there's all of this excitement about productivity and automation and new opportunities and new formats and all of that stuff. Marketers are very excited. And then on the other hand, you have Sam Altman talking about the need to create a system where we hopefully preserve some form of meaningful human work, with the implication being that we're heading toward a world with less best work for humans to do. And then you have Dario, the CEO of Anthropic, like, Scaring the Pants off of Everyone, talking about mass job disruption, the potential for misuse, authoritarian abuse. You know, it's very much like a, like a film, but it's reality. So when you zoom, or potential reality, when you zoom out from like media plans and ad budgets and tactical stuff and think about AI over the next, like, I don't know, it's. I really don't know what time horizon to give because things are moving so fast. But say five years, what's your gut feeling about where all of this is taking us? And is there anything about it that keeps you up at night or gives you pause?
- [46:04]
A
Such a big way to end this. Yeah, I mean, yes, social media, I keep going back to that, but there are parallels. In the early days of social media, it was thought of Very positively. These are new, exciting ways for people to communicate, reconnect with your friends. And it was very rose colored glasses. When you look back now on how social media was pitched as a consumer phenomenon and then also for advertisers and things changed. There's much more negativity surrounding what social media is and how it interacts and how it operates.
- [46:45]
B
It's having a big tobacco moment, for God's sake.
- [46:48]
A
Companies that, yeah, I mean, you know, countries around the world, they are banning social media use by young, young, young children and teens. They are, there's all, all these, still lots of questions about regulating these businesses, about the impact on society of these businesses. And so there's a lot that we, you know, in some, in hindsight, wish we would have learned earlier in thinking about the impact of social media. So I am heartened now by the fact that the founders of OpenAI and Anthropic are going out there and talking openly about the challenges, the problems, the issues. They're trying to get out in front of these things. Yes, you could say that these are marketing initiatives. At the same time they are cautionary, you know, phrases that they're saying, but they're opening up the dialogue and I think that's healthy. But I think the dialogue can't just be one sided. I mean, it does need to be something where industry, you know, marketers, agencies, consumers, regulators, government, everybody involved needs to keep a really close eye on what the impact of AI is. You know, we're talking about it in marketing, which is important, but it's not as weighty and consequential as the impact on jobs, on how people feel about themselves, on security, you know, all of these things. So while we're having a conversation about marketing and advertising, don't forget about all these other big things that does keep me up at night that does make me worry that is the AI really going to become so powerful that it does something that is very detrimental to society or humanity? And at the same time I promote the idea of AI media, I also go, is this a bad thing? I hope that the open dialogue that's starting is something that will help contain it and that these companies will recognize that if the AIs do become so powerful as they are predicting that they will be, that there will be a way to, to rein them in to, you know,
- [49:12]
B
Debbie.
- [49:12]
A
I know what a way to end a podcast, Allison.
- [49:17]
B
I mean, I, I really couldn't help myself, but I think that's a good message and I, I don't disagree that there's more awareness now if we're going to make the comparison to social media of potential misuses. At the beginning of social media, everyone was like, we're just going to bring the world closer together. And, like, the consequences were completely unaddressed. And it was totally Pollyanna. So, yes, we're already at the point where we're talking about Skynet moments, so I guess that's a positive thing. Let's just not reach the point where, I don't know, we're all. What was that meme that. Did you see? It was also a Black Mirror episode. It was like a bunch of. Of people on exercise bikes and they were like powering screens and that was how we got electricity into the world. And it was the only job left for people. And everyone was really buff and they were super healthy, but they had nothing else to do with their lives but sitting on these exercise machines.
- [50:15]
A
Oh, my goodness.
- [50:16]
B
Feeding energy into AI systems. Something like that.
- [50:19]
A
Yeah.
- [50:19]
B
But anyway, so let's not reach that. If we don't reach that, I think we're okay.
- [50:26]
A
Yeah.