This episode is brought to you by Novi, the infrastructure powering brand growth in AI commerce. By connecting brands, certification bodies, and major retailers, Novi ensures verified product data is accurate, consistent, and surfaced where shoppers and AI models search, turning credibility into authority, visibility, and conversion. Learn more at noviconnect.com
As we look ahead to the next holiday season, will your marketing strategy even matter if an AI agent is making the final recommendation for the consumer?
Agility requires more than just the latest AI tools. It sometimes requires fundamentally re-engineering how your brand earns visibility and trust in an algorithm-driven world. It demands a shift from winning clicks on a search page to becoming the definitive answer for an AI agent.
Today, we’re going to talk about how agentic AI is quietly becoming the new gatekeeper between brands and consumers, radically changing e-commerce discovery and purchase behavior, especially in the CPG and retail space.
To help me discuss this topic, I’d like to welcome, Kimberly Shenk, CEO at Novi, our Resident Expert on AI-Driven Commerce.
About Kimberly Shenk
Kimberly Shenk is co-founder and CEO of Novi, a technology company that helps CPG brands and retailers ensure consumers can easily discover and select their products when using AI assistants to shop. A serial tech entrepreneur, Shenk has led data science teams at early and midstage startups such as Eventbrite, Domino Data Labs and NakedPoppy, where she was a co-founder and Head of Product. Before transitioning to the private sector, Shenk served as a United States Air Force Captain for five years, holding the chief data scientist position at the Pacific Air Force headquarters in Hawaii. She holds a BS from the U.S. Air Force Academy and an MS in data science from the Massachusetts Institute of Technology (MIT). In 2025, she was named to the Inc. Female Founders 500 list for a second time.
Kimberly Shenk on LinkedIn: https://www.linkedin.com/in/kimberlyshenk/
Resources
Novi: https://www.noviconnect.com/
This episode is brought to you by Novi. Novi is the infrastructure powering brand growth in AI commerce. By connecting brands, certification bodies, and major retailers, Novi ensures verified product data is accurate, consistent, and surfaced where shoppers and AI models search, turning credibility into authority, visibility, and conversion. Learn more at noviconnect.com
Catch the future of e-commerce at eTail Palm Springs, Feb 23-26 in Palm Springs, CA. Go here for more details: https://etailwest.wbresearch.com/
Drive your customers to new horizons at the premier retail event of the year for Retail and Brand marketers. Learn more at CRMC 2026, June 1-3. https://www.thecrmc.com/
Enjoyed the show? Tell us more at and give us a rating so others can find the show at: https://ratethispodcast.com/agile
Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstrom
Don’t miss a thing: get the latest episodes, sign up for our newsletter and more: https://www.theagilebrand.show
Check out The Agile Brand Guide website with articles, insights, and Martechipedia, the wiki for marketing technology: https://www.agilebrandguide.com
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. https://www.missinglink.company
Transcript
[00:59:20] Greg Kihlstrom: As we look ahead to the next holiday season, will your marketing strategy even matter if an AI agent is making the final recommendation for the consumer? Agility requires more than just the latest AI tools, it sometimes requires fundamentally re-engineering how your brand earns visibility and trust in an algorithm driven world. It demands a shift from winning clicks on a search page to becoming the definitive answer for an AI agent. Today we’re going to talk about how Agentic AI is quietly becoming the new gatekeeper between brands and consumers, radically changing e-commerce discovery and purchase behavior, especially in the CPG and retail space. To help me discuss this topic, I’d like to welcome Kimberly Shenk, CEO at Novi, our resident expert on AI-driven commerce. Kimberly, welcome to the show.
[01:46:27] Kimberly Shenk: Thanks for having me, Greg. Excited to be here.
[01:48:65] Greg Kihlstrom: Yeah, looking forward to talking about this. I mean, definitely a, a timely topic here, so looking forward to talking about this. Before we do, though, why don’t we, uh, why don’t you give a little background on yourself and your role at Novi?
[02:00:66] Kimberly Shenk: Yeah, quick background on me. Um, I have been a data scientist my entire career. I studied at MIT back when AI was not a thing. I was a chief data scientist in the Air Force, but ultimately found my way into CPG after being at a couple of different tech startups leading how uh, consumers search and leverage data to power search engines. And um, that was kind of some of the foundational work that ended up helping me start and found Novi back in 2020, focused on CPG search and how brands stand out and show up for consumers.
[02:37:11] Greg Kihlstrom: Yeah, yeah. Well, and, and building on that, maybe you could talk a little bit more about, you know, what, what Novi does and your audience and, and, you know, what, who you primarily work with.
[02:48:26] Kimberly Shenk: Yeah. So we are uh, an infrastructure layer for brands to help them stand out and be visible to consumers wherever brands are selling their products. So in the early days of Novi, that was definitely search on the different major platforms. So think about Target and Ulta and all the different retailers where a brand would show up. And now that’s, you know, as consumers are shifting their search behavior to AI and AI platforms, it’s definitely the data infrastructure layer to help them show up in those platforms.
[03:20:17] Greg Kihlstrom: Yeah, got it. So yeah, let’s dive in here and I want to start with what I touched on in the intro is really how AI and, and other developments are really redefining brand discovery. And, and so starting from the, the strategic standpoint here, the, the concept of Agentic AI is moving us from a world of SEO and search results being the thing, you know, the way to, to discover to a world of direct answers. Strategically, what does this mean for a CPG brand that’s spent the last decade mastering SEO and SEM to win on the digital shelf?
[03:58:36] Kimberly Shenk: Yeah, it’s so interesting because we are living in a really crazy time where product discovery is just being fundamentally rewired. Like you said, last decade, brands won by ranking. And so that was the SEO and SEM game. And today they’re winning by being selected. And so that’s a totally different game. So the consumer is never at this point seeing a search results page anymore. The, they’re getting a synthesized answer from AI. So if your product data isn’t clear, it’s not consistent, it’s not siteable enough for AI to use, you’re just not showing up in that moment. And so, yeah, SEO, it was all about helping search engines know where to find your content. But AEO is about helping this AI engine understand your information well enough so that it can explain it in its own words, which I think is really interesting. That shift is very, very large. So brands that have built this like empire around SEM and SEO now have to retrain this muscle around not optimizing for keywords, optimizing for answers. So what’s interesting about this too is it’s leveling the playing field. AI is not rewarding the loudest brand or the biggest budget. It’s actually rewarding brands with the most trustworthy, structured, verifiable data. So smaller brands can beat category leaders if their data clarity is, is stronger. And so this is what we’re seeing is that so far for large brands, this strategically means that they’re exposed in a new discovery channel that consumers are, you know, more and more going to. And so the brand equity that they’ve spent millions of dollars building over the years really doesn’t isn’t in the same, where they’re not on the same playing field. So this is a huge risk, but this is also a big opportunity because smaller brands, they can actually gain massive market share. And so they’re, you know, if they’re agile and they are AI native, AI shopping native, they could be, this is a very strategic advantage for them.
[05:54:62] Greg Kihlstrom: Yeah, yeah. And I mean I, I like, I, I think this is a positive trend, you could say, and that, you know, it’s, it, we’ve all, we’ve been talking about high quality content being better, you know, for years, but at the same time, I think SEO would have eventually gotten there the way that it was, it was moving, but I feel like it’s now been accelerated in that, you know, to your point, the, the authentic, the trustworthy, the valuable content now becomes even, you know, even more important than it, than it ever was. Uh, but this also does make things more complex. You know, the, Novi put out a guide that we’ll refer to a few times and, and link to in the, in the show notes as well. And, and one, one of the mentions there is that AI makes the, the messy middle of the customer journey even more complex. You know, along that line, you know, how does a brand strategy for influencing consideration and evaluation need to change when an AI and not a human is, is doing most of the comparison shopping?
[06:55:04] Kimberly Shenk: Yeah. A lot of marketers have actually underestimated this one because that messy middle that you’re referring to, it used to be the, like the human opening up all the tabs and reading all the reviews and influencer posts and retail pages. And so, yeah, AI is compressing that. And so that entire mess messy middle step is the single step and it’s doing all the comparison shopping for the consumer. So really it’s evaluating three things that we, uh, we talk to brands a lot about. So first, like you mentioned, it’s trust. And so is the information verifiable? Is it consistent? Is it recognized by high authoritative sources? The second though is relevance. So is the information about the product available and it clearly maps to the intent of the question? So you just need to make sure you have a lot of information about your product so that when different questions come up, it can be clearly mapped and and it’s seen as a relevant product for that question. And then the last though is extractability what we talk about. The data has to be structured so the machine can actually read it. And so even if you have a lot of great content, we find brands missing out on, uh, showing up because it’s, they don’t have clean information. So if any of those are missing, you’re out of the consideration set. And so that’s what we talk about and I I’m constantly talking about is your product is no longer being merchandised to humans. It’s being merchandised to machines. They’re being vetted by a model. And so the models are very selective and so you, it means your job in this messy middle now is to pre-package data so that AI can actually find and confidently select select your information. So that happens with clean structured consistent claims, clear attributes and and um verified sources.
[08:36:58] Greg Kihlstrom: And I think that that part of optimization, I think we’re so used to optimizing for humans and, and what humans read. I mean, you, you mentioned several things there from a, from a data and and structure standpoint. But, you know, further to that point, what are maybe some of the things that are non-obvious or non-obvious types of data beyond product specs that might need to be optimized for AI agents to find and, and favor these products over others?
[09:04:19] Kimberly Shenk: Yeah, one big one and that’s non-obvious but absolutely critical is what we talk about verified claims. So AI weighs um, the verified but consistent sources far more heavily than promotional language. So when we talk about verified claims, think about all of the attributes about your product that are beyond product specs. So ingredient level disclosures or even certifications like USDA organic or FSC or testing results, things that give, you know, like efficacy claims, some backing, safety flags, environmental impact data. So there’s a a whole laundry list of things, but they are basically trust anchors or signals that AI can rely on. And then we talk about that in terms of consistency. Consistency is like if you have one product page at Target talking about a product as cruelty free, and then another maybe says not tested on animals at Walmart. Actually the AI doesn’t necessarily and always unify those things. And so that’s a claim, it could be product names. There’s a lot of things. So inconsistency is treated as uncertainty. And so that’s something that downgrades you. So you know, product claims are basically the, the thing that differentiates your product. So like we not product specs, it’s the layer deeper. And that is what AI is using to know if you’re relevant to the question. So like historically, everything marketers cared about on a product page was basic, like size and weight and color and a short product description. And the branding and the storytelling is what the consumer used to fill in the gap emotionally. And then of course, do research on reviews and dozens of blogs and influencers and what that. But AI is doing all that research like we talked about now. They’re compressing that messy middle. And so they just they have to have the data. And so that’s where actually the the un-obvious thing is enriched claims because that’s what differentiates your product and that’s what AI actually heavily relies on.
[11:04:47] Greg Kihlstrom: Yeah. And I, I want to get back to talk about keywords a little bit again because again, lots of focus on SEO and, and SEM over the years. It’s not like SEO has gone away, but, you know, with, with answer engine optimization, AEO, also sometimes called G-O, you know, pick your pick your acronym, right? With as as with all of these things. But a lot of the, the tools that are focused on AEO or or G-O are, are kind of continuing that keyword tradition, so to speak, because they’re focused on user’s prompts. But is that the right approach? I mean, you know, what, what should brands be focusing on if not?
[11:45:39] Kimberly Shenk: Yeah. So the blunt answer is no, focusing on keywords is not the right approach. It’s actually can hurt your AI visibility. And so the easiest way to explain this, our um, our head of AI research, he has this amazing metaphor. So I’m going to steal this from him, but SEO is like going and talking to the librarian. You ask, hey, where can I find a book to learn about communication? So that librarian points you to aisle three, shelf seven. And that’s what SEO does. It helps the system know where to find your content to go find it on the library shelf. But AEO is like actually turning to the research librarian and saying the same question, hey, where can I learn about communication? And they’re going to say, oh, Charles Duhigg teaches about great communicators. Listen first and then this other author talks about empathy and they’re going to explain the content in their own words because they deeply understand it. And so that’s the shift. SEO with with keywords, it was about organizing knowledge. So this keywords needed to be there so that that knowledge could be surfaced to a consumer and they could do their own research and interpretation on it. But AEO is doing the interpretation so the consumer doesn’t have to do any of that work. And then it goes even further to personalize it. And so this is where that prompt situation comes in is I might have a way different context and prompt history than you do. And I might talk to AI about a lot of different things and it’s interpreting things about me that are very different than it’s interpreting about you. So the people can go in and type the same exact prompt into AI and get completely different answers based on the context of what it knows about me versus you. And that’s obviously not the case in traditional SEO. You and I type the same thing, we get the same exact answer because everyone’s bidding on the keywords and we’re it’s just going to show up. So the success is being trusted enough by AI to be represented in the answer based on whoever’s asking. And so it’s not about keywords, it’s about, did you structure your data in an authoritative way clearly enough so that the AI could interpret it and then use it in its own words. So that’s the, that’s the heart of AEO and why prompt chasing doesn’t actually get you the the right outcomes that you’re looking for as a brand.
[03:57:42] (Music)
[14:00:81] Greg Kihlstrom: You know, we, we’ve touched on the, the, the platform as as well as some of the data components of this. But of course, this also impacts the people and the processes that, you know, that that they’ve been used to doing and they’ve been taught to, to use and, and all that. You know, how are you seeing marketing teams adapting anything from internal workflows to the skill sets needed to move from this, you know, SEO mindset to an AI optimization mindset?
[14:29:43] Kimberly Shenk: Yeah, that’s a good question. There’s a couple of things. So I’d say the first big thing that we see is marketing teams or organizations that have figured this out have accepted that they’re not just marketing to humans. They are still marketing to humans for sure, but there’s a packaging component for machines. So there’s workflows where teams are writing content that is a little bit more natural or scannable or structured or verifiable, free of fluff. So I think that’s like one area where marketing teams are starting to really figure out how to do better content production. There’s content and then though the other layer for CPG specifically, in terms of shopping, is understanding the data asset. It’s not just a technical chore anymore. It’s actually your competitive advantage. And so the the content is making sure that AI understands a lot about what you’re doing, but the product data is making sure that your product is actually being surfaced. And so you have to have cross-functional alignment between actually the R&D teams because the R&D teams that are developing the products know the accurate attributes. They understand how the product is made, what went into it. They can give you all that enriched information. Then the marketing team has to be able to take that with the positioning and the branding, but actually the e-commerce and the digital teams have to put that into the right schemas, into the right feeds, make sure the data is integrated. And so that’s a new skill set. And so finding folks that are really in up to that, um, and treating data as a a core advantage. And then I think the other thing honestly that we’ve seen though from a more of a philosophical perspective is that it’s a advantage for early movers. And so in CPG, a lot of these retai and brand and retailers are are a little behind. And so we’re seeing the ones that are ahead are recognizing that traditional SEO, SEM, and even paid ads in other channels like Meta are just not returning as much as they used to. And instead, they’re, so instead of thinking of AEO as incremental budget, they’re actually reallocating budget. So they’re taking that bottom part of the ad spend and part of their SEO spend and recognizing, oh, we’re not getting as much traffic from SEO anymore. We’re going to reallocate it to AEO. And so those are the ones that are winning versus the ones that are thinking of this as incremental. Because if you think about it as incremental, then you have to go back and advocate for more budget. That requires very clear ROI metrics. You have to build a business case. And here you are stuck in this like internal minutia and then you’re behind. And so reallocating budget allows you to do stuff this quarter, not wait for clarity, and then that’s how you show up. So like we’re seeing teams that are reallocating budget now are showing up in 40 to 60% of AI responses in their category versus the ones that are trying to think about us an incremental project. So yeah, those are those are the the top three things I I’d say we’re seeing.
[17:11:51] Greg Kihlstrom: Well, and speaking of of metrics and, and measurements, I mean, certainly, you know, down down the funnel, the the metrics are probably similar, but, you know, looking at things like click through rates, search rankings, you know, a lot of this stuff just doesn’t apply with with some of these approaches. So, you know, are there new KPIs or measurements or, you know, what how should how should marketing leaders be thinking about measuring success here?
[17:37:66] Kimberly Shenk: Yeah, it’s a good question. We’re seeing the AEO journey has really three different measurement phases. And the first one is about readiness and it seems a little bit obvious, but a lot of it is like is your data right now even consistent and structured enough for AI to see it? And so we see marketing teams starting out with just an AEO site audit and then um product description page consistency scores. So it’s basically telling you is your data even visible and interpretable by by models, by LLMs? Have you removed all the blockers that keep AI from trusting or using your product information at the start. Without any of this, none of the other work matters and so this is like obvious but most important phase. Then the next phase is more about momentum. So that’s when your data is AI ready and you’re actually measuring if you’re being surfaced. So this is share of voice is the North star. And so it’s the share of voice is just how often you appear in AI answers for ones that you should be showing up for, for your category and for your product type. And then the percent of products that you have that are ranking in these engines. And so this is telling you, okay, AI systems are starting to cite you, they’re including you, they’re reusing your product data. Early signal that AEO is working. And then the last phase, which is now where like you’ve done that early work and this is the steady state over time is consistent visibility. So momentum’s built, but you need to make sure that your appearances aren’t sporadic. You want them to be like predictable, repeatable, all of those things. And so this is share of voice is still the main metric, but it’s making sure it’s stable over time. You’re not being tested by AI anymore. You’ve just now become this trusted, reliable product source that it’s pulling into answers consistently.
[19:21:84] Greg Kihlstrom: Yeah, yeah. So another thing here, and, and you mentioned that, you know, being, being part of the the results in, in AEO is, is a component of this. But, you know, when, when AI consolidates 10 sources into a single recommendation, how do you how do you do marketing attribution in, in those cases? You know, how can how can a brand prove ROI of their AEO efforts when their brand name might just appear in a summary without a direct click?
[19:51:30] Kimberly Shenk: Yeah, this is the hottest topic and something we talk to every single brand about. So, so we already talked about, you got to measure obviously readiness, did you clean the data? Momentum, are you showing up and then visibility, consistency, is is it stable over time? But like you said, you can’t measure clicks. And so you have to measure how the visibility is increasing or translating into traffic. And so traffic is, are more people coming to you directly because you were recognized in AI. And so you should see this translate into three types of traffic and this is where we see most of the ROI calculations coming from today. So first is branded search traffic. And so when an AI engine is repeatedly surfacing your products, consumers will start directly searching for your brand and your products, your SKUs. So this is one of the clearest indicators that we’ve seen that AEO is actually influencing real world real world demand. The second though is direct traffic. So this is the I already know what I want, take me straight there behavior. Um, this is actually going to increase when consumers see your brand and products repeatedly in AI recommendations. So it is another early signal that becom you’re becoming more of a default choice uh, for AI and then for new consumers. But then the third, which is really interesting is direct referral traffic. And so even though AI is reducing the number of clicks, there’s this moment where they are still going to cite your brand and they’re going to give a link. And so you’ll see spikes in this direct referral traffic that’s going to it’s honestly the best correlation or correspond to you successfully being placed in the answers. So, yeah, we think about it obviously as like readiness, momentum, visibility. So share of voice is still really important, but then traffic is really the measurement of ROI and attribution today.
[21:37:34] Greg Kihlstrom: Yeah, definitely. And, and, yeah, so, you know, just having been through the the holiday shopping season here, you know, what what did you see from a consumer behavior standpoint that maybe caught brands off guard or what maybe some brands at least caught off guard, um, that that ended up happening with with AI shopping?
[22:00:54] Kimberly Shenk: Yeah, I think it’s so interesting especially because um, even ChatGPT like five days before Black Friday released the shopping research functionality. And so what we really saw was consumers relying on these AI assistants for curation and doing the research on their behalf, which seems obvious, but like they went and the AI selected the products, compared the prices, checked availability. And in some cases, you know, we’re starting to see building a cart essentially. But I think what brands haven’t fully internalized yet and the thing that was really catching caught a lot of us off guard to be quite honest is AI becoming this category manager if you will. So for example, I mean we all have trusted the big box retailers for their ability to manage the category and do curation for us. You know, that’s why you walk into Target. They have sowered the earth for the best, most compelling products and they’re curating them for you, so you trust Target. But now the consumer is shifting their trust to AI to do the research and curation. And so I don’t think that’s the thing that brands and retailers are really ready for. And it’s going to be really interesting to see that grow and that consumer behavior um, deepen in the research world.
[23:14:83] Greg Kihlstrom: Yeah, yeah. And, you know, going, going beyond transactions and, and things like that, how do you, how do you see AI altering the long-term relationship and even customer loyalty loop between CPG brands and, and their customers? You know, can AI actually help build a stronger, more direct connection or is it always going to be this kind of intermediary at best?
[23:40:53] Kimberly Shenk: This is a really interesting question cuz I think a lot of loyalty also comes from the benefits a retailer or brand can provide. So things like efficiency, ease, more savings like from a loyalty program. So you might get a for example, like a recommendation for a beauty product in AI, but people are die hard for their loyalty program at Sephora. So then you might just still go straight to Sephora to buy it because of the loyalty program. Or I mean think about it like this, same thing with Amazon. You might get a recommendation, but you really want to use Prime. So you’re going to go find it on Amazon and then buy it. So that I think is we we still have a lot to learn and we’re going to have to watch how consumers behave. But I will say that no matter what, I think AI will actually strengthen long-term brand loyalty, especially for the ones that embrace this shift because um, again, loyalty is built in AI is built on trust and transparency. And so it’s not based on legacy brand equity. And so the real opportunity is making sure that AI remembers. So once you are consistently selected, models repeat selection. It becomes this really durable memory. A loyalty loop of the future if you will. And so it’s just this interesting like if you’re investing in that data, AI mediated loyalty is actually likely going to be stickier than historical loyalty that was just built through advertising.
[25:00:84] Greg Kihlstrom: Yeah, I love that. Well, um, as as we wrap up here, a couple, couple things for you. If we were having this interview one year from now, what is one thing that we would definitely be talking about?
[25:12:44] Kimberly Shenk: Ah, I think that we will be talking about something even bigger than AI assisted shopping, but fully agentic commerce. This is going to be crazy, but uh, shopping agents won’t just recommend the products. They’ll do the full end-to-end research, compare, decide, and then transact on behalf of the customer. So I, I think it’s going to be very interesting um, to see when the agent is handling the entire process.
[25:38:81] Greg Kihlstrom: Yeah.
[25:39:56] Kimberly Shenk: It’s going to be, you know, like the consumer isn’t doing any part of it. Uh, the emphasis of data clarity and the machine needs that. And so I think that’s one thing that we’re going to be talking about and uh when they’re building the cart, the agents are choosing the brands, the agents are reordering on behalf of the consumer. So that’ll be really fun to see. We obviously all have predictions about it, but I think we’ll definitely be talking about it in a year.
[26:04:16] Greg Kihlstrom: Yeah, I love it. Yeah, let’s, uh, let’s, let’s definitely let’s get you back on and we’ll we’ll we’ll talk about that. Love it. Well, uh, Kimberly, thanks so much for joining today. Really appreciate you, um, sharing your ideas and insights. One last question for you before we go, what do you do to stay agile in your role and how do you find a way to do it consistently?
[26:22:98] Kimberly Shenk: I like this question. So two things. I think first is staying on top of the real consumer behavior. So staying deeply plugged in, how are people actually shopping, not what we think they’re going to do. And then treating learning as a daily habit. So every shift comes down to understanding and discovering patterns early. I have a GPT I created that sends me the most relevant news on consumer shopping and model updates every morning, running experiments consistently, staying close to brands. So I guess for me, agility is really just like curiosity that I practice consistently. So yeah. But thanks so much for having me on. I really appreciate it.
