The Mid-Market Brand Trap No One Is Talking About in the Age of AEO
Once AEO works for everyone, the cluster becomes the problem most teams haven't priced in.
Three things shipped in 21 days: Visa launched payment rails for AI agents, HubSpot launched the first CRM-embedded AEO product, and Shopify quietly switched on agentic storefronts for eligible stores. The agentic commerce stack stopped being a future conversation, and most B2B teams are still arguing about whether AEO is worth the spend.
This episode walks through what is actually inside the retrieval pool right now. A live demo across Perplexity, ChatGPT, and Claude shows what an agent reads when someone asks for a CRM recommendation: aggregator content dominates, and brand-owned pages do not get cited. That is where the argument starts.
The real problem starts after AEO works. Once every brand in a category gets technically retrievable, the agent has to pick from a cluster of similar content embedded in the same vector space, and familiarity wins by default. If you are a series B through series D company about to drop six figures on AEO this year, this episode is about the mistake your category is about to make at scale.
Where to jump in
- 00:00Why this episode and why now
- 01:43Visa, HubSpot, Shopify: what just shipped in 21 days
- 04:19Live demo: what Perplexity, ChatGPT, and Claude actually read
- 07:56Why AEO succeeding is the trap
- 13:26What actually moves a brand out of the cluster
Why agent bias is the mechanism the dual-journey paper names
This episode is downstream of the working paper. The dual-journey model argues that B2B brands now serve two audiences in sequence — the AI agent that retrieves and shortlists, then the human buyer who decides. AEO determines whether you make it into the retrieval pool, and agent bias determines whether you survive the shortlist that comes after. This episode walks through the second half of that argument with a live demo and the clustering mechanism behind it.
Working Paper · SSRNRead the conversation
You or your CMO are about to spend about six figures on AEO this year, and I'm going to show you why that spend actually accelerates this little problem that we have, and why the infrastructure that just shipped about 3 weeks ago is actually making this a whole lot worse.
Why this episode and why now
Welcome back to the 360° Marketing Show. I'm Anusha. I research how AI is changing marketing, not just with what's trending this week, but actually going deep into the mechanism to understand what this could look like in the future. The reason I'm talking about this this week and not next quarter is because of what happened in the last 21 days that has kind of made me think a lot about AEO because it is fundamentally changing what AEO actually means.
Things like Visa shipped payment rails for AI agents. HubSpot shipped this first CRM embedded AEO product. Shopify turned on agentic storefronts by default. The agentic commerce stack became a production system this past month. And every B2B marketing team that I know, thought leadership post that I read are having a completely different version of this conversation that seems a bit off to me.
Here's what I'm covering. First, the three dates from the last 3 weeks and what they actually unlocked. Second, a live demo across Perplexity, ChatGPT, and Claude that shows what the agent is reading right now. Let's not even call it agent, but we'll get more into that later because it's almost certainly not your brand. Third, why HubSpot AEO and every AEO tool coming behind it is structurally the productivity trap wearing a retrieval costume. If you walk away with one thing from this episode, I want it to be the shape of the mistake your category is going to make, not a better playbook for shipping into it.
Three weeks of agentic infrastructure
First, let's talk about the three weeks that's actually changing this infrastructure. The week after my birthday, April 8th, 2026, Visa launches Intelligent Commerce Connect. It's a payment infrastructure that lets AI agents purchase things on a customer's behalf. You don't have to enter your card details anymore. It literally has it. And it does this with spend controls, tokenizations, and fraud authentication built into it. And apparently, they also have pilot partners that are already running end-to-end transactions. Visa's own data says 47% of US shoppers already use AI tools for at least one shopping task. These buyers that are actually using agents right now are actually going to hand off decision-making, purchasing, etc. completely in the next 18 months.
On April 14, 2026, one of my favorite companies, HubSpot, launches HubSpot AEO at their Spring Spotlight. First major CRM embedded AEO product. From what I see, I feel like it uses your CRM data to surface prompts that your customers would ask in LLMs. It tracks how your brands show up in things like ChatGPT, Perplexity, Claude, Gemini, and optimizes it for that accordingly. HubSpot's own numbers on this are: organic traffic to their customer sites fell 27% year-over-year. AI referral traffic tripled. That is the curve that they're pricing this product into.
And something else that is already live and is actually rolling out really silently is Shopify's Agentic Storefronts. Apparently it is active for eligible stores right now. I don't use Shopify, let me know if you do, let me know what your thoughts are on that. Because it covers ChatGPT, Google AI Mode, Gemini, Microsoft Copilot, and it makes your product catalog machine readable to agents. I just came in here to tell you to watch this video because it talks a lot about machine readable content. But let's get one of the things that I noticed: if you're on Shopify, you are already in the retrieval pool whether you opted in or not. It's kind of like how LinkedIn just started to use our data without telling us — like we figured it out cuz we're smart, but the same way.
So, do you actually see the pattern? Three different layers — payments, marketing, commerce — all shipping almost the same thing in 3 weeks. The agentic commerce stack is not a future conversation. It is happening right now. But let's be honest, I'm not a journalist and I'm not here to talk about what top three things that happened this week. I'm more curious about what is happening underneath all of this. You look at all of them, what they do is the same infrastructure. All they're doing is putting brand content into the retrieval pool that AI agents can read. That is the shared direction and that is literally what AEO means.
Live demo: what the retrieval pool actually contains
Speaking of retrieval, let me actually show you what is happening in the retrieval pool today. I hope the answer to that is going to change your mind about how you think about AEO spend.
What I'm going to do is run the same query on three different tools with the same question, which is — I have it copy paste — "What's the best CRM for a series B startup with a, say, 10% sales statement." Now let's do the same thing.
Now look at where it is actually looking, which is very interesting for me. I'm betting HubSpot, okay, before I can even talk about it. So, you see HubSpot, you see Salesforce, I bet we'll see PipeDrive as well, Attio... interesting. That also says Attio is interesting — it says it's become the CRM of choice for AI native companies. None of these companies are sponsoring me. But what I'm really interested in is seeing the sources that it's actually opened.
It's looking at all these kinds of blogs, right? Do you see anywhere? Same thing, do you see that? Now let's go into this one. That's what I told you. Salesforce — actually, retrieval search bars, search engines are literally doing the same thing. Look at the kind of pages that it's actually looking at. "Best CRM from..." — is looking at SalesFlare, LinkedIn editors, Zapier, TechRadar, Sendes, Polo, Sendes. Again, not a single piece of — no, when I look at this, not a single piece of this seems to be brand owned. Do you see the brand name showing up in any of these? I don't.
If you actually look at this, the most interesting part about all these searches for me is that HubSpot is the only one that is actually showing up with its own citation. Right? This is super real. Now this is the argument: HubSpot is the only brand in this response with enough authority and aggregator density to occasionally get its own content pulled in. PipeDrive hasn't gotten it. I don't see PipeDrive. I don't see even Salesforce. I don't even see Attio.
So if you really think about it — what is happening is that agents are failing to differentiate between brand content. They're not reading brand content at all. Your blog, your landing pages — they aren't in the conversation. The G2 summary of the box, the Zapier comparison of you to the Zap... Look at this. There's a Zapier comparison. The 12 best CRMs for small business app — comparison of you to 12 other tools that are literally in the same space as your category. Now this is the retrieval pool of your brand as it currently exists. It is essentially a summary of your brand by people who may or may not have lost your point.
So this is what AEO is working towards, right? Because this is not exactly an agent. This is a retrieval mechanism. It's a visibility problem where your own content is not being surfaced. So the AI retrieval mechanism is actually looking for aggregated sources that it can pull, that get enough volume to justify and form a semantic layer and be like, "this actually makes more sense for what the user is asking."
That is why companies like HubSpot and Salesforce will always win, because they dominate the market. They have the most rev volume, most comparison data. There's so much volume about the company and also frankly they're the most familiar names in the training data. Familiarity does dominate when the pool is aggregator only.
AEO as a productivity trap
So, AEO, Shopify Agentic Storefront, every single one of these tools, it exists in the category to change that, which is to get your content into the pool. And right now, that is the most useful thing you could do to get your brand out there in front of people. I'm not here to argue that AEO is fake or that it's not needed. I'm making a very different argument.
Here's the argument. So let's imagine that AEO has actually started working for everybody. Just like SEO, it has started working. Everybody has the right tools. Everybody knows how to surface and be cited on ChatGPT, Perplexity, Claude, etc. And once your content is in the pool, the selection mechanism is actually going to change. The very tools that help you get in front of people are going to be the reason that you can't really do what is to be done, because you don't know what is to be done anymore.
See, the more we talk so much about AEO because it makes us feel so productive. We love to talk about things that are trending, which is AEO, because it makes us feel productive, like we're actually making a difference. But I think AEO is just a productivity trap wearing a retrieval layer outfit. Because the fact that AEO is actually succeeding, it means that the agent is actually reading your blogs, your landing pages, your product descriptions, your customer stories, but it is also reading every single competitor's content the same way.
We all belong in the same cluster. It's like CRM — HubSpot, Salesforce. "Oh, they're saying the customers are this, they're saying the customers are this. Oh, I find reviews about this, I find reviews about this." Do you see the problem? It's not just that the human buyer is going to have this confusion, but an agent — agentic tool — is also going to have it, except that it doesn't have feelings. The question stops being about who's retrievable and becomes "wait, who's actually distinguishable?"
How clustering works inside the model
I wrote a working paper on SSRN about the dual journey that I think we're about to take, because agents are inherently biased. And I also have a video on that if you want to check it out.
When AI agents have to pick from the same pool of similar content, the technical term is called context matching, which is a huge conversation right now. People are only context engineering, not prompt engineering anymore. There are great companies out there doing great work with context engineering. You should look them up.
I'm about to put on my technical hat. Just bear with me for a second cuz this gets interesting. All right. So what happens now is, the agent embeds each piece of content into a high-dimensional vector space. Embeds the user's query into the same space. It's going to try and pull the closest matches.
It's kind of like walking into a room filled with all of my girlfriends. Like, I've had so many girl best friends, but I only pick the ones that are really resonating with the things I'm going through and really understand what I'm trying to say. Doesn't mean the other best friends are not cool. It's just that these two people understand me a lot better right now than the other best friends do. Do you see that meaning? And that's how ranking seems to be working inside the model.
The part that I haven't seen people talk about in AEO conversation is that — if every brand in a certain category has been producing similar kinds of content, frameworks, how-to guides, webinars, same AI-assisted drafts; and if every brand uses the same AEO tool to surface CRM-derived prompts; and if every brand feeds that tool similar customer data pulled from similar public sources — all of those brands embed into the same region of the vector space. They essentially cluster.
And the reason I'm talking about this is because I was recently thinking about a solution to solve one of these problems. And I thought about how I would go about pitching this to companies. And then I stopped myself. I was like — I'm not sure if you're going to pitch the same solution to different companies that are competing with each other. You're essentially going to put them in a cluster and they're going to fight to get out of that or to find some other solution. It's not really — there's not really a green solutions, marketing-oriented. However, there are certain things that can really work for everybody.
The agent now looks at that cluster and is really confused about how to differentiate them. Then it just goes to the secondary stuff, like pricing, you know, validity, SLAs, stuff like that. That is not what marketing is working towards, because the essential — like the ultimate goal is to make sure that all brand becomes the preference, not just the price points. Now when you take these things, it becomes brand familiarity, which is already available in the training data, and you're just essentially going with the same thing you're seeing today with an aggregator, just with a different name for it.
This is what I mean when I say AEO is a productivity trap wearing a retrieval costume. AEO is going to produce interchangeable content inside the retrieval pool by making the same kind of optimization target — you know, things, same prompts, schema, answer engine surface — and the operating target across each competitor in the same category, the same mechanism. It really is the same mechanism, just with a different surface.
Who AEO actually helps and who it traps
So let me be precise about who this will actually affect. Because if you are a big brand and you have a huge hold on the market, like if you're HubSpot, Salesforce, Amazon, Google, AEO probably works really well for you because you're already a familiar name. AEO just makes you more findable.
And if you're the smallest brand in a category, like you're just, you know, you have enough signals to make you be like the differentiator, the up-and-coming brand, AEO will actually also work for you because you go from invisible to at least in a pool — like you at least belong to a cluster.
The brands that really would have a problem are the ones that are in the middle, like series B through series D. Legitimate products with real customers that actually have, you know, good amount of history, but they don't have content that is articulating these great points about the products and companies and customers. Now those brands will spend a lot of money only to get clustered with the same brands, you know, that have spent — that have used the same AEO tools, same money, but they get a different result in a doomsday.
What moves a brand out of the cluster
What is the actual solution? What moves you out of the cluster is not another tool. I'm not going to recommend trying another tool. It has never been — it's the thing that actually got me into marketing. Kotler talked about it in 1967 and Bezos talked about it in 1996. What every marketing text has been saying for the last 60 years, which is specificity, knowledge of your customer, no competitor has expressed in content that reflects it.
See, just like how a viewer can tell if a video is completely scripted by AI versus someone actually having thought this through or making a point — like in this video you may be able to tell, this is completely things that I have thought about — the same way the agents can also tell the difference between a brand that knows its customer and a brand that knows its category. The agent is not attending your town halls where you're talking about what's best to do for your customers. It's not reading your internal memos if you're not feeding it. But it can read what you have publicly.
I think I'm actually — I'm actually going to stop there, because the next layer, which is what you can do operationally, how you rebuild a content operating system or content operation around this specific reality that has not dawned upon us yet, but will soon — I think it's a longer conversation. It's in a book I'm researching right now, which I might need your help on. So if you have any take on this whatsoever, feel free to leave a comment or message me on LinkedIn.
What I want from you
So if you are a senior marketer or if you have a CMO that's about to spend a couple million dollars on AEO, it is going to do one useful thing. It's going to put your content on the retrieval layer that agents will actually read right now. That is not an entire waste of time or money, but it is also not essentially going to solve the problems that we may see coming in the next couple of months or maybe next year. Because once you're in the pool and everybody else is like your competitor, it's going to get crowded.
The thing the pool has everything to do with is not really AEO, but what you actually know about your customer and how well you're able to convey that to the world. The brands that figured this out in 2024 are going to look really well evolved in 2027. But the ones that are running their first AEO program as their strategy this year are going to look like marketing teams that went all in on "pivot to video" — which is strategically correct but just one layer technically behind.
If you are a senior marketer, CMO, or founder and you're thinking, okay, what do I do that is not this AEO strategy — honestly, do the AEO strategy. Just, there's something else that we could talk about. So message me on LinkedIn and I'd love to talk about that, because I'm writing a book and I'd love to get your thoughts on them.
Two things. Subscribe to our newsletter, because this launches a day before the episode and it actually goes deeper into things that I probably wouldn't be able to cover in the video. If you are the marketer making AEO decisions at your company, I would love to include your take on the research for my book. For more information, link in the description, and I'll see you in the next one.
Are you a senior marketer thinking about agent-era strategy at your company?
The research behind the second book is open. Senior marketers, CMOs, and founders shaping how their organizations adapt to agent-mediated buying — your perspective is what the book is built on.
Apply to be interviewed →