Subscribe
Share
Share
Embed
10 minutes of expert insights every weekday. Your morning ritual for staying ahead in retail media.
AI Advertising: Retail Perspectives and Predictions
===
[00:00:00] Kiri: The conventional wisdom says that AI advertising is a future problem, something to monitor, maybe something to experiment, but something that doesn't really demand immediate attention. But Debbie Ajo Williamson would disagree. [00:00:15] Debbie is the founder and chief analyst of Sonata Insights
[00:00:18] and publishes the AI Ad Economy newsletter on Substack. Before that. She spent 19 years at eMarketer and was one of the first analysts to identify social media as something that marketers should pay [00:00:30] attention to. Writing her very first report on it In 2006 when Facebook was just two years old, she now brings that same pattern recognition lens to AI's impact on advertising and commerce.
[00:00:43] Today I'm gonna share some [00:00:45] snippets from a conversation that she had with Scott Wino on the Retail Gentech Podcast, where they covered the chat GBT ad rollout, the competitive AI advertising landscape, and what retailers doing in response and where all of this is [00:01:00] headed.
[00:01:00] Let's jump in
[00:01:01]
[00:01:02] Kiri: Highlight number one. This is not a test. Chat. GBT launched its ad test about 30 days before this episode was recorded. And Debbie is realistic about what it actually is. Let's [00:01:15] listen.
[00:01:15] Debbie: this is a test that's leading towards something much bigger. Uh, chat, GPT is hiring for ad.
[00:01:23] Engineers, product managers, developers, um, people to lead the ad business. So, [00:01:30] um, this is not something that they're going into lightly or as an experiment to just kind of see if maybe advertising fits. Uh, this is definitely the start of. Where advertising might be going within chat [00:01:45] GBT. Um, it is early, right?
[00:01:47] And I think that what, there's a lot of things that are, are going to change over the next couple of months as we start to see some results from these early tests. Uh, one thing that [00:02:00] I know that chat GBT and OpenAI have been leaning really heavily into is the idea of trust. And that means the ads are separate.
[00:02:11] The ads are clearly identified as. Ads, [00:02:15] um, the ads don't impact what people see in the responses that they get. I mean, Chachi, BT has been very, very careful in all of its framing around this launch to, to make that very clear. Uh. [00:02:30] It's important, right? I mean, you think about it, I mean, you don't want, I mean, you've built this sort of new economy, this new environment where people are going to get information.
[00:02:39]
[00:02:40] Kiri: She also noted that the current ad formats are deliberately basic. [00:02:45] They look like search ads because that probably made it easier for advertisers to launch into the program and compare performance against existing search campaigns.
[00:02:54] Don't mistake simple for unambitious. This is scaffolding not the finished [00:03:00] building. Number two, 18 years versus 18 months. One of the most useful frames from this conversation was Debbie's comparison between Amazon's ad trajectory and open ais. Amazon existed as an online [00:03:15] retailer for roughly 18 years before building an advertising business.
[00:03:20] Chat GPT, while not a retailer, got there in a fraction of the time.
[00:03:26] Debbie: Amazon existed for a really long time before [00:03:30] advertising and now. Amazon is one of the largest sellers of digital advertising in the world, right?
[00:03:36] I mean, you've got Google, you've got Meta of course ahead of that. But if you look at Amazon, I mean, it's right up there. So it's been able to successfully [00:03:45] build an advertising business. Um, and people are still using Amazon. People are not like going, ah, ads. Too many ads Maybe. Maybe they're running away a little bit, but they're still using Amazon.
[00:03:57] Che, GPT obviously started [00:04:00] incorporating advertising way sooner, right? Um, but. To me, I mean, everything about AI is moving faster than it ever moved for Amazon. Um, back in the day, I mean, it made sense why Amazon took a long time to incorporate [00:04:15] advertising because it was building a business in a different era than we are now.
[00:04:20] And so the Adverti, the fact that O OpenAI already has advertising, I think just tells you about the speed at which this transformation is taking. Place [00:04:30] in how people are discovering, finding, researching, and eventually buying things. And I don't think OpenAI could afford to wait. A year, two years, five years.
[00:04:40] Um, they needed to move F fast because they knew what they had built in [00:04:45] terms of something that could, that would be valuable to people. And then I think they realized that this was also going to be a really valuable audience and environment for advertisers.
[00:04:54]
[00:04:56] Kiri: The speed comparison matters for retail media too. Debbie [00:05:00] and Scott discuss how retailers are among the heaviest advertisers. On chat, GPT Sensor Tower data showed roughly a quarter of early advertisers were in the retail category, including Target Williams Sonoma.
[00:05:14] Best [00:05:15] Buy and Albertsons. Debbie called this the frenemy conversation. Retailers want to learn from these platforms, but they're also afraid of being left out of the referral equation. Amazon and Walmart meanwhile have [00:05:30] their own AI assistance that protect their existing retail media businesses, Which is why formats like sponsored prompts inside Rufuss and Sparky are getting attention as potentially AI native ad [00:05:45] units. Miracle Ads is the Ad Tech solution trusted by Rakuten and over 50 global enterprise retailers. That's because Miracle Ads [00:06:00] was built with both three P Marketplace sellers and one P suppliers in mind. Both advertiser audiences demand a seamless advertising journey from onboarding to reporting.
[00:06:11] Kiri Masters: You can offer everything from sponsored products to [00:06:15] video ads all in one solution. Learn more@miracle.com. That's M-I-R-A-K l.com.
[00:06:24] Kiri: Number three, ads aimed at agents, not humans. The [00:06:30] conversation closed with Debbie. Looking ahead, and this is where it gets interesting for anyone thinking about where retail media and AI advertising converge.
[00:06:39] Let's listen.
[00:06:40] Debbie: I think that, um, chat GPT is not always going to be just like a [00:06:45] basic little search box in the middle of the page. There's gonna be different ways that you interact and engage with chat GPT depending on what you're doing, you know, whether it's shopping, whether it's looking for entertainment, whether you're planning travel, uh, they're going to be different interfaces and that will open up the opportunity [00:07:00] to have different types of ads.
[00:07:01] Formats, um, more akin to the spectrum of ads that advertisers are used to video ads. I know ads look like social media ads in environments when are a little bit more social. Um, you know, ad ads that lead to, you know, per direct [00:07:15] purchase opportunities. So I feel like, you know, we're really just at the beginning stages of what advertising is going to look like.
[00:07:21] Um. I don't, I expect there'll be brand new formats that we have never heard of before, uh, that advertisers are going to have to get used to. We saw that [00:07:30] with social media. Uh, we where, where Facebook companies like Facebook came out with ads that looked just like content in the feed. And you had to, you had to figure out, well, how do I as an advertiser show up?
[00:07:40] Um, you know, among all of those feed posts that I, that, that, that are in [00:07:45] someone's social media feed. Uh, they did figure that out and it became hugely successful. So I feel like at AI we're going to see formats that we've never even comprehended before. Uh, and then, you know, honestly we talk, there's so much talk about agents and agentic [00:08:00] commerce, and I think that is a shift we're going to see more and more development in over the next few months and into 2027, uh, buying.
[00:08:11] The ability to have an agent go out and buy for you, um, [00:08:15] to make a transaction for you is going to start to become more of a reality. And, you know, not to get too, too fu futuristic, but it's going to lead to an environment where the advertising is not aimed at the human eyeballs, um, or the human [00:08:30] sensibilities at all, but is aimed at the agent and is designed to get the agent to perform an action.
[00:08:36] So it's not even going to look like advertising. We might start to see glimmers of that in the next year.
[00:08:41]
[00:08:42] Kiri: Now that last point connects to something I've [00:08:45] been covering closely. The Columbia and Yale research I wrote about last year showed AI agents already penalized sponsored products by eight to 14% in selection probability. Today's agents have learned [00:09:00] to avoid ads entirely. But if Debbie's right, and I think she might be the next generation of AI advertising won't really look like ads at all.
[00:09:09] It'll be structured data, product feeds, trust signals designed to [00:09:15] influence agent decision making rather than necessarily catching a human eye. Now what? Debbie draws on nearly two decades of watching consumer behavior shifts turn into advertising economies.
[00:09:29] She tracks [00:09:30] social media from a billion dollar ad category to the foundation of Meta's empire. Her rate on AI advertising is that we're in month one of a transformation that will play out over the next. Decade and the pace that we're [00:09:45] seeing just in these early few weeks makes social media's evolution look leisurely
[00:09:53] For retail media specifically. The question is about how retailers are going to react to this. We have retailers that are already [00:10:00] testing these ad units on chat, GBT. We have retailers who are building their own apps inside the chat, GBT marketplace.
[00:10:09] We have retailers who are focused on the foundational work of building structured product data and [00:10:15] thinking about how their onsite ad economics might change.
[00:10:20] Maybe some of them are hoping that this is all a bad dream and just suddenly goes away, but it won't we'll link up to Debbie's newsletter, the AI ad Economy on [00:10:30] Substack, and listen to the full conversation that Debbie and Scott had on the Retail Gentech podcast.
[00:10:36] Thanks for listening, and I'll catch you next week.
[00:10:39]