Latest / Retail Media Breakfast Club / AI Agents in Retail Media: Why Media Buyers Aren’t Ready for the Agentic Revolution
Transcript
- Media buyers are nowhere near ready for the agentic revolution === [00:00:00] Kiri Masters: AI agents. Everyone and their mother seems to be launching them. That's doubly true in retail media, where a buyer might be running campaigns across a dozen networks, and a good number of those are mandatory [00:00:15] under the terms of their supplier agreements and JBPs. The pitch kind of writes itself. Now, keeping track of the announcements is a full-time job. [00:00:26] Management platforms like Sky and Pacvue, [00:00:30] agencies, and even retailers themselves are launching agents designed to optimize every stage of media planning, buying, and reporting. An army of agents running your ad campaigns for [00:00:45] you? How can we possibly fill all this free time? The natural next question amid all this promised efficiency is who gets cut out? [00:00:56] Does the agent displace the platform, the [00:01:00] agency, or even the retailer's own sales team? That's a column for another week. A new survey from Sky, published in July, proposes a duller question first: Is anyone [00:01:15] actually using this stuff? [00:01:17] Today I'm sharing a piece that I published to my column at The Drum on August 4 Named Media Buyers Are Nowhere Near Ready for the Agentic Revolution Let's jump in. [00:01:29] [00:01:30] Kiri Masters: So first, the research. [00:01:32] Skai is a ad management platform, and they surveyed [00:01:35] 332 paid media practitioners in April this year, scoring them across nine dimensions of [00:01:45] agentic readiness [00:01:47] And the mean score across product categories and levels of seniority was 35.7 out of 100 [00:01:56] So using this distribution, Sky broke [00:02:00] out the respondents into these five levels of agentic readiness, ranging from unprepared to exploring, to building, to operational, to fully agentic. And spoiler alert, no one [00:02:15] really reached that agentic level just yet But 30% sat in unprepared [00:02:22] And 49% sat in exploring levels one and two out of five [00:02:29] Just [00:02:30] 3% are at the level that Skye calls operational [00:02:33] Which among other things is when the media buying team has AI embedded at scale in routine workflows [00:02:42] So there's different categories or [00:02:45] industries that were covered off in this report. Retail and e-commerce made up about 35% of the sample and scored with an overall readiness score of 37.4, which is slightly above the [00:03:00] cohort mean, but still below [00:03:02] A readiness level of three, the building stage of readiness. CPG and FMCG media buyers specifically scored 32.2, which [00:03:15] was the lowest of all six industries that Sky broke out [00:03:20] Where this gets interesting is in how Skye splits out seniority in the report. Agentic readiness scores slide [00:03:30] steadily down the career ladder. C-suite respondents self-report an agentic readiness score of forty-one point seven. Meanwhile, the managers the folks with their hands on the keyboards report [00:03:45] thirty-four point six. [00:03:47] And the interpretation from Skye in their comments in the report is that the leaders are the ones describing the strategy that is being presented in the board deck, while the [00:04:00] practitioners are describing the disjointed systems that they actually log into. Costco isn't scaling [00:04:15] its RMN with legacy ad tech and manual solutions. It's leaning into a cloud-centric composable stack. As part of the Costco Velocity Network, GrowthLoop empowers the [00:04:30] retail media team to build audiences directly from the Data Cloud, enabling faster activation, greater relevance, and better performance, all while maintaining privacy and governance [00:04:45] standards. [00:04:46] The result? Exceptional value and experience for Costco members and better ROI for brand advertisers. Learn why the GrowthLoop Composable Commerce Media Solution was the right choice for [00:05:00] Costco's Velocity Network. Visit go.growthloop.com/breakfast. That's go.growthloop.com/breakfast [00:05:15] Now, could it be that some practitioners are quietly sandbagging the technology that might redefine their jobs? [00:05:23] This report suggests otherwise, or at least suggests that they're not doing so consciously [00:05:28] People at least [00:05:30] in this survey seem to feel fine about agents in the abstract [00:05:35] Human and change readiness towards agentic automation is the strongest dimension at forty-six point five, while psychological [00:05:45] safety scores sixty-two point three. The likelier story here is worse and kind of less interesting. The leadership sees what this technology can [00:06:00] conceivably do, but the teams are the ones that are still stuck dealing with what it really takes to run a client campaign across a dozen Byzantine platforms [00:06:11] ultimately, agentic media buying really ought to [00:06:15] favor retail media. Retail media offers real-time sales data, conversion signals, this closed-loop view of transactions. So an agent that is optimizing towards units [00:06:30] sold rather than just clicks should really regard retail media as one of the most legible channels of them all [00:06:37] Now, one thing that's interesting to note from this data is that adoption of agentic media buying [00:06:45] appears inversely correlated with sophistication. Tail advertisers, meaning brands with smaller budgets, use algorithmic buying because they have no alternative. Sophisticated head advertisers [00:07:00] refuse to use it because they have their own capabilities and little interest in handing them over. [00:07:06] The mid-tier, the actual growth market for a lot of retailers, got burned by platforms like Google's Performance [00:07:15] Max in its early days and is understandably wary of anything resembling a black box algorithmic buying platform Pet retailer Chewy has built for this tension. Frank Mulcahy, who leads [00:07:30] Chewy Ads, described the design to me as a barbell, an objective-based product for vendors without the time or team to manage bids across twenty-five networks, and a clean room for the dozen or so [00:07:45] multinational CPG companies that were always going to run their own numbers anyway [00:07:51] There's a version of the next eighteen months in which every network, agency, and platform independently builds its own [00:08:00] agent to roughly the same specification because that is the current flavor of technological one-upmanship. If everyone gets there on more or less the same timeline, the first mover advantage evaporates, [00:08:15] and all that remains is the token bill [00:08:18] But the announcements will keep coming. Every network, platform, and agency will ship some kind of thing that automates media buying. Meanwhile, [00:08:30] the people who would have to use it are still exporting CSVs, and the retail and CPG teams funding this industry aren't exactly leaning in. Perhaps someone should give those eager [00:08:45] C-suite executives a DSP seat of their own [00:08:49]