Why AEO is Overhyped

So we're talking about SEO, we might as well talk about why I think AEO is overhyped. Because if you look at the Ahrefs, Ahrefs just posted yesterday, SEO company, ChatGPT traffic is flat in November. And you look at Gemini 3, oh, they're starting to kind of take their users a little bit.

Are you talking about ChatGPT referral traffic or the traffic for chatgpt.com?

On what the first one is traffic to chatgpt.com. The second piece is referral. I don't think referral traffic has changed that much. So yeah.

So when you're saying ChatGPT traffic is flat from Ahrefs, they're measuring the traffic going to chatgpt.com?

Correct. Yeah. Okay. Yeah. And so what, like, I'm not saying don't do AEO. I think we get a lot of people asking us for AEO. But the thing is, I think it's not even 1% of traffic for ChatGPT. It's 0.25% of web traffic is ChatGPT. And I'm like, and I remember talking to, you know, a pretty significant company a couple weeks ago. And like, the reason is like, dude, if it's less than 1%, why is everyone just focusing on this brand new thing over here when search traffic is still 42%? We're not saying search traffic is going to remain that way forever. But I do think right now people are overallocating and the game is actually search everywhere optimization, not just like, not just, you know, AEO. Yeah.

If you knew the data I knew for ChatGPT, you would be shocked.

Shocked in a good way or a bad way?

Bad way.

So like even less?

Yeah, I signed an NDA. It's one of the largest corporations in the world.

And we'll just leave it at this is overhyped, yes?

I don't think it's actually overhyped. I think it's here to stay. It's very powerful. I'm not talking about just the technology, I'm also talking about it as a marketing channel. I think it's just early innings and people haven't figured out how to really monetize it.

When I say overhype, hype is like for the short term in my mind. So it's hyped, it's overhyped in short term. I'm not saying it's not going to work. We said this with AI in the beginning too, in general. Yeah.

I think AEO is underhyped in the long run and how few companies are focusing on it when they should.

Yep.

I think the revenue expectations from it in the short run are too high, but I think in the long run they're underestimating it because it doesn't matter who wins. You and I know this is the new format. You need to do SEO, AEO, whatever you want to call it, AI SEO, in the future to do well, no matter which platform is going to win. And I think the platforms is not going to be winner takes all. I think it's going to be multiple players owning the space for different circumstances.

Yep.

And the best example I have for this is when I talked about the results for one of the campaigns, and I'm telling you, you would really be shocked. I was in this company's office and They're asking me which one is going to win, okay? ChatGPT or Google. And I'm like, they're both going to end up taking it. They're like, I don't see it. I only use ChatGPT. So then I asked them a question. What social network do you use? What guests do you— if you had to pick one, what one do you think they would pick?

So like in what context?

Just generally use, like every day. I'll give you the answer.

YouTube?

They said Instagram.

For me, it's YouTube.

Okay. And they're like, I'm like, okay. So There's a lot of social networks out there. Your social network's Instagram, right? I'm like, do you have a TikTok account? I do, but I rarely ever log in. How often do you log in? I don't know, once every month or 2. Okay. Do you have a LinkedIn account? Yeah. How often do you log in? A few times a month.

Mm-hmm.

Do you have a Facebook account? I do. How often do you log in? Maybe only once a week. All right. Do you use YouTube? Yeah, I use it every day. So I'm like, there's not one winner. You use it for different things. Do you use X? Yeah, I use X every single day.

Mm-hmm.

Right? So I was like, so you're telling me you only use YouTube, but in reality, are you saying that YouTube's your social network, but you use a lot of social networks? And then I start talking about use cases versus Instagram versus YouTube. And they're talking about YouTube's a little bit different because YouTube offers things like makeup tutorials that Instagram doesn't. And I was like, You sure it really offers different types of content? I'm like, I bet you that makeup tutorial that you're looking for on YouTube is on Instagram. They're like, I don't know. So then I'm like, okay, show me something that you watch on YouTube. They show me it. And guess what? Within 6 minutes, I was able to find that video on Instagram. Instagram, their search feature sucks compared to YouTube. But I tried searching for it, couldn't do it, then ended up looking up the creator of it, had to scroll a lot, found that video from the creator. To make your point. To make my point.

Yep.

And the point I was getting at to her, I was just like, you use different social networks for different things. X, she used it for news. People will use ChatGPT and Google and Gemini and Perplexity and all these other solutions. And I really look at Gemini as Google because most people that I know don't go to Gemini directly, but a lot of their stuff is integrated into Google. I'm assuming you agree with that.

Yep.

And I believe the social networks are going to have— I mean, the AI platforms, LLMs, will have different use cases from Perplexity to ChatGPT to Google and so forth, so on. Just like how people have different use cases for social networks, even though they have access to a lot of the same information and content.

Here's the problem with most marketing agencies. They don't understand what is happening with the world of AI. With my agency, Single Grain, we think about agentic workflows. We think about creating agents for the future. When we think about SEO, we think about creating different products such as internal linking products or content consolidation products. And on the other side of things, we think about building products such as Kare. We just think about making marketing easier overall. If you want to stay at the cutting edge of marketing and at the same time leverage the power of AI, go to singlegrain.com and we'll see you on the other side. So 2 things here. We should talk about some of our use cases for the different LLM platforms because you and I pay for everything. And then I want to talk about how Goldman Sachs talks about the real ROI from companies like JP Morgan, Chipotle, Target, and UPS. But So use cases. So real AI use cases I have from the different LLMs. ChatGPT is my go-to because it has memory, right? And so I store everything in there. I can think strategically or remember what I talked about with, you know, people we need to recruit or processes, things like that. Okay. But what do I use Claude for, right? So Claude I use for code.

Code.

Yeah. So code, when I want to analyze the data we have. So I've been doing this recently this week. I was like, hey, look at all the closed won deals from HubSpot and tell me, look at all the Gong transcripts tied to that and tell me what people are talking about and using the Ahrefs MCP, what keywords we should be going after, okay? Also with the closed lost deals, look at all the gone transcripts. What are people's objections? What should we be optimizing for? Oh, also do a win-loss pattern to show me what the, what, like, for example, for our won deals, we have 34 touchpoints, okay? For our lost deals, it's 60% less. It's like, okay, that's some stuff that, you know, normally you can't get.

Well, are you basing the lost deals based on when you lose it? Because if you look at the lost deals and if you lose it early on, you're not going to have the same touchpoints.

I don't want to touch on that. I'm making a point. So, so, so, But so going to— so we use Claude, we use Gemini, I use it for VO3, I use it for image generation, I use it for the Gemini Gems because you can pull from YouTube, I use it for NotebookLM. So my point is making your point too. It's like we have different use cases for these things. I'm curious, what are your use cases for the different ones?

So I can break it down into 2 simple buckets. For personal stuff, I use ChatGPT. For business stuff, I use Google.

That's easy. Yeah.

It's really easy because Google has more data. And we— I find in business, uh, like if I'm asking tax-related stuff or corporate structures, Google tends to give me better responses than what I get from ChatGPT. They're similar, but sometimes ChatGPT, I found this to be a little bit more inaccurate, which is too risky on those kinds of things. Personal stuff, I'm asking things like health, you know, sleeping better, neck pain, stuff like that. And I use it because of the memory.

Yep.

So then that way I don't have to keep going back and starting over.

And I like Perplexity, it's not necessarily an LLM because it uses the other ones, but I like it for deep research. It's really good for stock research and I want different charts and things like that and also like projections and all that, so.

I've tried it for that. I've tried Grok for that and I've tried Google and ChatGPT. I found them all to be off when I'm trying to do that.

I just want it to be directionally correct, so.

Yeah, I want it to be 100% accurate. That's the problem because if I'm going to make a trade based on the data, I don't want anyone sending me—

I want to feel like 80% good about it. But like Grok, I use Grok quite a bit. Like if someone tweets something with data, I just want to know, I want to hit the button to like verify it. So I'll use that. So there's different use cases for everything. But going back, Goldman Sachs reveals real AR—

AI.

Goldman Sachs reveals real AI ROI from companies like JPMorgan, Chipotle, Target, and UPS. So you and I, we've been talking about like what's the true ROI, right? So I'll start here. So let's start with JPMorgan. So they use AI to reduce headcount. growth while revenue climbs. So Bank of New York, that's not that specific, but they, Bank of New York Mellon, they built 117 AI solutions with 100 digital employees. And then Rocket Mortgage implemented AI to cut loan processing time from 4 hours to 15 minutes. It's not bad. Okay, let's go to Target. Did I say Target? Target uses AI for forecasting across the organization. Home Depot implements AI across supply chain and stores. And then Walmart built 4 super agents for customers, employees, and advertisers. And then you have UPS, which saved 38 million liters of fuel. That's pretty hefty, right? AI processes 90% of cross-border transactions. And then J.P. Hunt automated 100,000 hours of work across operations. So we're starting to see real ROI. Chipotle, huge improvements in kitchen automation, right? So I think the savings, maybe it's not quite apparent yet, but I think we're starting to see it.

I think we're really gonna start seeing the savings going into 2027.

Yeah.

I think companies are starting to implement it.

Yeah.

And when you look at what they're implementing, it's still a small fraction of their business. I think we're really going to start seeing it in 2027 because I know people are worried about the deteriorating job market, which in theory the CME Watch Group's predicting what, like an 89% chance? Is it 89% that there's a Fed rate cut?

Yeah.

Let's see, 87.2%. And I really look at it from a perspective of You're going to start seeing the impact more in 2027. I think there's also going to be new jobs created and it's not going to be as catastrophic as people are thinking, just like we've seen in past with history. I could be wrong, you know, I'm not the smartest person out there, but I do believe people are going to just have to learn new skills. It's going to impact jobs. They're going to have to learn new skills and start doing new things.

Let's talk about this piece because on one end, the Elon podcast I heard, I could just, let's just take a small detour for fun. If you were to pick 2 stocks right now, not financial advice, guys, what would you pick?

Yeah, again, not financial advice. My 2 favorites—

Because we did this a couple months ago.

Right now, Google. Google's the easy one for me. The number 2 is a toss-up for me, and I'll tell you why. It would either be Microsoft or Netflix. Okay.

Wow.

Netflix, because I think they've really won that market when it comes to streaming content.

Bought Warner Brothers today.

Yeah, I don't know if that's going to go through, but whether they buy that deal or not, I really do think Netflix has won that mindshare from the perspective of you just stream content there and people are really hooked on it. It's a new version of TV. The reason it's a toss-up between that and Microsoft is With Microsoft, they really have their grip. I would actually, I'll pick Microsoft as number 2.

Okay.

They have their grip on the corporate ecosystem, and I don't see that going away anytime soon. The reason I am reluctant about Microsoft is their products aren't that great for a lot of them, but we tend to use them because of the—

Bundling.

Bundling and the pricing. But some of their products like Teams isn't as good as Slack.

Yeah.

But I'm okay with it because it's, A lot cheaper than paying for Slack and Zoom.

Yeah, I would go Google number one, because you and I know Google well. I mean, the data that they have, the gem, like they're just like they're on a rampage, a tear right now. The TPUs, all these things, right? Waymo, all these things. I would go Tesla number 2 because of robotics. And I think that's if he gets it right, like Elon gets it right, that's going to be huge. Interestingly enough, when Elon did that interview with the other guy, he doesn't actually hold anything. He doesn't invest. He's like, I just like to build things and build on top of that. But he's like, if I were to pick, it would be Google, because of how far they are with AI right now. And then his number 2 was Nvidia, which is interesting. So anyway, this leads to the final topic that we have.

Wait, let's go back really quickly. You talked about Elon. I'm assuming you love it heavily because of robots and the future is going there. I just watched an interview with Boston Robotics, Boston Dynamics CEO. Boston Dynamics used to be owned by Google at one point, then it got bought by SoftBank. Now it's 80% owned by Hyundai and then 20% by SoftBank.

Interesting.

Okay. Did you know when Elon, He mentioned this on CNBC. Did you know when Elon's been doing the showcasing of the Tesla robot, a lot of it is powered by humans?

I didn't know that.

Yeah. And did you know the Boston Dynamics versions are not powered by humans?

That's okay.

They're much further along.

That's okay. So I don't think it's going to be winner take all, but I would go Tesla.

I agree with you. And did you know like Boston Dynamics already has tons of robots in warehouses and corporations that are leading the pack? I'm not saying they're going to win. I agree with you. It's just they don't have the cachet as let's say Tesla does.

Yeah, I would also go Tesla too because of robotaxis, right? So robotaxis, robotics, all that stuff. So anyway, let's move to the last topic here. So this is the final one. So this is, it's always good for us to talk about things where we have like a, I might have had an argument with somebody earlier this week, right? Like I feel strongly about this, but my take on this is no, your job is not safe from AI. Like, so one of the debates I had was like, oh, You know, but my job is safe because I'm going to be talking to people all the time and the human-to-human connection. Absolutely, I agree with that. Human-to-human connection is important. But would I rather take Neil with human-to-human connection or Neil with human connection plus AI? Like, it's equivalent. It's no different to me. Like, if you say, it's clear to me that if you don't want to learn AI, you're just saying to me, Neil, I don't want to come to work with a laptop. I don't want to come to work using the internet today. I don't want to use electricity. That is what it's the equivalent of. And I don't think 99% of jobs, I think, are like, I'd rather pick someone with AI versus not. And one of the debates that we were having was just like, oh, but like, no, you know, we have other things to focus on right now. Like, so should we wait like 2 more years, 3 years? How long should we wait?

It's a terrible way to think. I think those kind of people are the ones that are going to get passed by by others and other corporations.

Yeah. And then it's kind of like, they'll kind of just throw it to the side and joke like offhandedly, like as if nothing's going to happen. But Again, can I work for you if I don't know how to use the internet?

Nah, you can't.

Yeah.

Do I think you need to be the best person with AI? No, but as a human, you need to try to improve yourself and continually learn more and get better and adapt with where the future is going.

Now here's a question, philosophical one. How— curious how you guys are doing it at NP Digital, but what level of training are you going to do for AI for your people?

I don't know. That's a question I have to ask our CEO. My co-founder and I are really different.

Yeah.

We tend to just hire the people that already have good knowledge and already using it because I don't want to have to tell people, hey, do this and do that. I want people who are bringing something to the table and be like, oh, I've been doing this cool stuff that you guys haven't even seen yet. You should check this out. Because they're like, Self-motivated. Those are the people I'm hiring. I'm not really trying to hire people that I had to convince to use AI and train them up. I'm hiring more so people right now that are good to go and they're bringing stuff to the company and teaching our organization new ways we could be using AI that we never thought about that they're using already.

So the way I see this is we want to hire the best of the best people, right? We want to hire people that have high agency who are biased to action. Well, the thing with me is like, It's never been easier to find the people who have the bias to action because before it's like, you have to say, oh, what are you learning? Can you show me some case studies, whatever? Now it's just like, hey, what are you doing with AI? So what would you do for us? How much are you spending per month out of pocket? Some of the people I interview, $500 a month. Someone we talked to earlier, $2,000 a month. Okay. So I'm not saying you need to spend $2,000 a month, but it's very clear, like if we want the best of the best, then that should be one of the first things that we look at. And then one of the pushbacks that I get is like, oh, well, How are we going to find these people? Like, they're not available everywhere. Of course they're not. High agency bias to action people are in short supply always.

Yes. And you got to go hunt them out. They tend not to apply for jobs.

Exactly. Exactly. What's the winning formula for 2026?

There is no winning formula. Think of it as cooking spaghetti. Try new channels and experimentation. Be open-minded. Take the spaghetti, throw it against the wall. Whatever sticks usually means it's cooked. So think of your marketing that way. Throw it against the wall, whatever's starting to get traction, double down on that. Whatever's not, push it off to the side and get back to it later. But you got to keep experimenting, trying new things, and you got to do it at a faster clip than you were doing it in 2025 or even 2024, just because AI is causing people to move so fast. Look at these venture-funded companies now. People now expect the norm to be $0 to $30 million or $50 million in revenue in like 5 months, 6 months. That is not normal.

Yeah.

But they're assuming that, and that's all because of AI.

Yeah. All right guys, that's it for today. Catch you later. Bye. Bye.