What CPG Brands Actually Control with Agentic Commerce Optimization (ACO)

Agent Commerce Optimization for Consumer Brands
===

[00:00:00] Kiri Masters: A lot of consumer brands are already investing in GEO or generative engine optimization to make sure that their products are surfacing. When someone asks an LLM for recommendations, box [00:00:15] checked, moving on. But while these GEO approaches might get you found in AI conversations, it doesn't necessarily mean that your product is ready for the next era of [00:00:30] commerce, which is agents who shop and transact for buyers.

[00:00:34] When AI agents actually start transacting completing purchases on behalf of consumers, they're gonna need something different. They will need [00:00:45] canonical product data, structured attributes, rich context, and that's what Scott Wino from Retail Gentech and refi buy calls a agentic commerce optimization.[00:01:00]

[00:01:00] And the checklist looks a little bit different to what we're seeing in terms of best practices for GEO.

[00:01:07] But the bigger problem for CPG brands who primarily sell through other retailers and [00:01:15] don't do a whole lot of volume through their own D to C site. Most of that checklist isn't yours to control. The fate of your brand.

[00:01:24] When it comes to AgTech commerce is really in the hands of your retailer partners. So I'm [00:01:30] going to talk through today the difference between GEO and agen commerce optimization and and from a consumer brand perspective, what is in your control versus what is in retailer's [00:01:45] control and what you should be asking them about.

[00:01:48] If you're a retailer listening to this, I highly recommend that you check out the original agent commerce optimization framework at Retail Gentech. I'll link up to it in the show notes. [00:02:00] That's going to be more relevant for you. You are the enabler or blocker for your brand suppliers on everything that I'm about to describe.

[00:02:09] Let's jump in.

[00:02:10]

[00:02:12] Kiri Masters: So first, the difference between GEO and a [00:02:15] CO. So GEO is getting you discovered in LLMs today. When someone asks perplexity or claw or chart, GPT for dog food recommendations, your brand might show up because [00:02:30] of brand recognition. Lots of SEO work that you've done in the past, maybe even some recent GEO work that you've done in the recent past, and so that's great visibility.

[00:02:41] A CO gets you chosen [00:02:45] and available for purchase.

[00:02:47] It ensures that when an AI agent is ready to transact, it has everything it needs to show your product as a purchasable option within [00:03:00] the LLM or within the browser, whatever you're using.

[00:03:04] It is the difference between Purina cat food appearing in search results when I'm asking about what type of cat food I should buy for my 9-year-old cat and the agent [00:03:15] actually recommending a specific Purina cat food skew. Knowing that it's in stock at the nearest retailer that I buy from, understands that it's grain free and then is able to actually complete the purchase.

[00:03:29] [00:03:30] So if you're a CPG brand that sells primarily through retail partners rather than your own website, you don't get to control some very important foundational requirements here like. Crawlability on your website, and this is a [00:03:45] really big one for retailers because just by default a lot of these retailer websites have turned off the ability for agents to crawl their site by default, and it's just, you know, it's a security thing from the past.

[00:03:59] [00:04:00] They're not allowing agents to crawl their website. There's nothing you can really do about that as a brand, right. Besides being educated and maybe asking your retailers about it and impressing on them, maybe they should be thinking about it. [00:04:15] Miracle Ads is the Ad Tech solution trusted by Rakuten and over 50 global enterprise retailers. That's because Miracle Ads was built with both [00:04:30] three P Marketplace sellers and one P suppliers in mind. Both advertiser audiences demand a seamless advertising journey from onboarding to reporting.

[00:04:40] You can offer everything from sponsored products to video ads all [00:04:45] in one solution. Learn more@miracle.com. That's M-I-R-A-K l.com.

[00:04:53] So there's a few more other steps in this nine step framework that you don't have [00:05:00] direct control over.

[00:05:01] And those include, the infrastructure. So is the site accessible to agents? Number two is crawlability, is the catalog accessible to engines and agents? You don't get to decide [00:05:15] either of those two things, but steps number three through seven, you do get to partner with your retail partners on. So those are basic product content.

[00:05:26] This is the usual stuff. Titles, images, pricing, [00:05:30] description, variations is step number four. ~Number five is E-commerce attributes.~

[00:05:35] ~Step ~Step number five is basic e-commerce attributes. So this is category, features, specifications. These are all things that you have [00:05:45] likely submitted to a marketplace in the past,

[00:05:49] But the retailer themselves controls. The final implementation of that. How does it look onsite once they have it in their system? Step number six is [00:06:00] expanded attributes, and this is where we're starting to get into things that are a little more specific to ai, so more context that an agent.

[00:06:11] Needs about user [00:06:15] reviews. What have other customers said about this product in the past? Cross sell. What products do customers often buy this with, or after researching this category, which products do they often buy? Usage [00:06:30] context. How is this used? Who uses it and when? So brands can.

[00:06:36] Potentially provide some of this contextual information to retailers, but retailers are going to [00:06:45] decide how that is displayed and provided to an agent. Step number seven, and this is the final piece of information that a brand could conceivably work with their retailer partners on, [00:07:00] is extra product content, which includes reviews.

[00:07:04] Q and as user generated content and multi-lingual content as well. So brands can pass this content to the [00:07:15] retailer. The retailer, again, decides how that is aggregated and and displayed to an agent. So in all of these cases. The brands provide the inputs, but the retailers control the outputs. [00:07:30] There are two other steps in the A CO framework where brands can and should really be tracking how this works.

[00:07:37] And it's essentially the modern version of. Digital shelf, tracking and optimization. So [00:07:45] step number eight is product card monitoring. So our product card is how your product shows up in an LLM or how it shows up in a browser as a search result. [00:08:00] Essentially

[00:08:00] like how today you would search on Amazon. For a certain product and you'd get the list of product results. A product card is sort of the AI version of that. It shows up with an image, a title, pricing, et cetera, which [00:08:15] retailer it's available from. So retailers are obviously gonna want to have product cards in those search results, but brands will also wanna be tracking.

[00:08:27] If and where they're showing up for certain queries. [00:08:30] So shelf analytics vendors should be starting to evolve their offerings for a agent commerce, um, scenario. So that is something as a brand you'll wanna be on the lookout for, you know, is, is your [00:08:45] current digital shelf vendor moving in that direction?

[00:08:48] Are they starting to track results in LLS?

[00:08:52] So what this means for brands is you can create perfect agent ready product data, but if your retail partner doesn't [00:09:00] implement it correctly or at all, it's not going to matter. So this is where brands can start to have conversations with their retailer partners about what is their plan for this?

[00:09:13] What can you as a [00:09:15] brand be providing to your retailer partners to help them move along this journey? To be clear, this is going to be a huge lift for retailers. There's a lot more infrastructure required for this is,

[00:09:29] but [00:09:30] as your distribution partner.

[00:09:33] It is certainly within reason to want to understand where they're at in the journey and how they're planning to optimize for this scenario. Some of these concepts [00:09:45] a little harder to talk through in an audio format. If you wanna see that full list of nine steps, I suggest check out the blog posts that accompanies this episode for a table and.

[00:09:56] If you wanna dive a little bit deeper, certainly recommend checking out the [00:10:00] Retail Gentech blog, which talks through all of this, certainly from the perspective of a retailer or a brand that has a really strong D two C site.

[00:10:10] But it really is in your interest to understand what your [00:10:15] retail partners should be building and how the partnership needs to evolve.

[00:10:19]