Latest / Retail Media Breakfast Club / AI Retail Media Buying: Are We Handing the Keys to the Bots Too Soon?
Transcript
- Bots, Automation, and Friction in Retail Ads === [00:00:00] Kiri Masters: So I've been clearing out my house on Facebook Marketplace lately, and I noticed something pretty ironic that seems to be new. So Marketplace has long let buyers fire off [00:00:15] an instant message for items that they're interested in, and it says, "Hi, is this available?" And that message goes out with one tap. So anyone who's sold on there knows the exhaustion of getting flooded with these [00:00:30] messages that are all the same, "Hi, is this available? [00:00:32] Hi, is this available?" So what is Facebook's latest fix to this conundrum? Well, they gave sellers an auto-reply feature. Now my automated [00:00:45] responses answer buyers' automated inquiries all natively in the platform. So Facebook created this problem, then solved it by automating both sides of the conversation. [00:00:59] So we have [00:01:00] bots talking to bots and nobody actually deciding anything at all, [00:01:05] [00:01:07] Kiri Masters: which is kind of like the pitch that some retailers are making to advertisers right now: [00:01:15] objective-based, AI-driven ad buying. Tell the platform your goal, hand over the keys, let it optimize the bids, let it optimize the placements, the products, the channels. [00:01:27] And retailers are racing to launch [00:01:30] these. I've profiled a number of them in my columns over the last six-plus months. But I kept wondering whether the advertisers themselves actually want this. So I asked. [00:01:45] I ran a poll on LinkedIn, 139 people voted on it, and here's what they said. So there was four answers to this question of how do you feel about objective-based, [00:02:00] AI-driven retail media buying? [00:02:02] And 12% of people said, "Bring it on. I want automation." 41% of people said, "Interested if it's transparent." 37% said, "Skeptical, I want [00:02:15] control." And 10% said, "Hard pass." So the bulk of those answers are right there in the middle, interested if it's transparent or skeptical, they want control. That's four out of five respondents objecting to handing [00:02:30] it over blind. [00:02:31] Now is kind of sharing the same instinct that I might use this, I might be open to it, I wanna see what's under the hood first [00:02:41] Not an outright pass, at least for only ten percent [00:02:45] of people object to it completely in principle. And there are a few great comments on here that I wanna share with you. I am planning a longer, more in-depth piece a- analyzing the different sides of this question for my column at The Drum. [00:02:59] But I [00:03:00] wanna do a little preview. I had some great comments from folks on LinkedIn that I wanna share. And if you've got some thoughts on this, I would love to hear from you as well while I put this all together. [00:03:15] The agentic AI revolution is here, and the competitive edge belongs to those who move faster than disruption itself. Join Miracle Ads and me in New York City on [00:03:30] June 10, 2026, for strategies, insights, and connections built for the era of agentic commerce. Discovery is shifting, media economics are following, and the brands and [00:03:45] retailers who figure out what comes next won't be the ones who wait. [00:03:49] They'll be the ones already moving. Join me at the Miracle Summit in New York City on June 10. Link to register in the show notes [00:04:00] So the first comment was from Corey Buller, who is the senior director of commerce media at Dentsu, and he said that this was the vendor Epsilon's entire off-site targeting strategy across their retail media partners and always performed [00:04:15] badly on incrementality. [00:04:17] His reasoning is that the AI is built to find the most efficient attributable returns, so it generally over-indexes on people who were going to buy anyway, and the automation optimizes towards [00:04:30] the sale that didn't really need any help. Second standout comment was from Evan Walsh, who leads partnerships at the measurement firm Incremental. [00:04:40] He named two hurdles that he keeps hitting. One is the retailer [00:04:45] bias. Without the right controls and visibility, buyers stay skeptical about what is being prioritized. Two, what he called ROAS jail. These systems need a primary KPI to organize against or to optimize [00:05:00] against, [00:05:00] and often it is not to drive total sales, it is to improve ad metrics. So if you pick any base ad metric, you're right back to where you started. And finally, James Tenser, longtime retail tech analyst and journalist, raised [00:05:15] the question that sits above all of that: What if the bots only optimize for short-term outcomes? [00:05:21] Where does brand equity go? If retail media wants a permanent seat in the marketing mix, he argued, buying cannot be confined to [00:05:30] narrow technical objectives. So the resistance, at least from these folks who took the time to comment, is less about whether the machine can do the job. [00:05:41] There didn't seem to be a lot of doubt as to the technical [00:05:45] capability of these algorithmic buying models, but it's more about whether you can see what it's doing, and above all, a concern that advertisers might be pulled in to optimize something that [00:05:58] shouldn't be optimized [00:06:00] at the expense of everything else. So like I said, I'm chewing on this one a little bit for a longer piece that I'm planning. I shared these poll results with the head of Chewy's ad business in an interview last week, and he had some great [00:06:15] comments to make. So a lot more on that soon. [00:06:18] My question to you is: when we automate both sides of the transaction, are we removing the friction or are we just removing ourselves? Thanks for listening. I'll catch you [00:06:30] tomorrow. [00:06:30]