The 2-Year Window: How Mid-Market Companies Can Win the AI Search Race | AI For The C-Suite EP 70

I'm Chad Harvey and this is AI for the C-suite. The show for senior leaders who know AI matters and need to figure out what to do about it. Each episode we dig into what it all actually means for middle market organizations, how it's changing decisions, strategy, leadership, and the nature of work. Let's get into it.

My guest today has been at the intersection of AI and search since 2008, well before the current wave. From 2019 to 2025, he was embedded at Ford as what he calls an intrapreneur, architecting their tier three digital and AI transformation. His work drove a 700 % increase in buyer intent traffic and shifted share of voice from 26 % to 60 % in Ford's top markets. Early in his career, he built the FBI's first social behavioral inference model, a real-time gang chatter detection dashboard now used nationwide.

He holds two patents, helped launch the Lexus AI Music Lab during Lexus's strongest sales run in 35 years, I guess a year to remember as the commercials go, and recently authored a book called Maintaining Visibility, a strategic framework for GEO in the AI search era. Today, he runs e04.ai where he helps brands stop being invisible to chat GPT, perplexity, and the new generative layer of search. And if you stretch the definition of neighbor generously, He's also a Bucks County PA neighbor of mine.

Alan, welcome to AI for the C-suite. Chad, thank you so much for having me. Thanks for the great introduction as well. Absolutely Alan, I am really excited for our conversation today because there has been so much change in development in uh the world of AI and whether you want to call it generative engine optimization search or anything else.

There's a whole lot of terms out there right now for the same thing. This is going to be a great episode because I get a ton of questions about this from folks and usually they start from someplace like this. uh I talked to an exec. And they express some type of feeling to me like they should be doing something about AI search, but they can't actually articulate what's changing beneath their feet and what they should do.

So given that this is your wheelhouse, let's start here from your perspective. What is the shift that's happening right now? So let's talk about that and why is it different from a normal SEO update site? Sure, and we have 60 minutes, right?

Because I can go on a little further. So, I mean, in the basic, basic sense, what's happening is you now have, and we're at AI assistance about to be AI agents. um the shift on the human side is almost 60 % of people are now using AI. for the big purchases.

And that covers SaaS in terms of this audience and how they compare it. And that's the big shift. You can still get, this is the company, this is what they do. A lot of that's still available through AI.

You're not invisible yet. the rest of it, though, um if you have subject matter experts and they're building authority for your product or trust, They're invisible. They're not connected. Their entities are not connected to your product.

So when someone wants to compare one product over another, that is very difficult. And we saw the SaaS apocalypse. And a big reason, even with that, some of the data that came out, was people were unable to do that deeper level comparison, which is your first step. Yeah, I'm visible.

If you want to check, you're probably visible. the rest of the pieces of it that you rely on for the sales, a lot of that's not there. Okay, I think one of the things that I get a lot of pushback on from folks that have even just a couple gray hairs in their head is that I've been through these cycles before, you know, we change first there was mobile first and then there were featured snippet shifts a decade or so ago. So how is this different?

How is this shift different right now? So in the old days, um Google produced 10 links. It went out and it found them. Today, AI is relying on what it's learned.

And we don't know exactly when it learned those pieces. And that's one of the reasons we see so much Reddit now in the results is because of uh the rag technology and what that's able to do. in real time to validate they're using um Reddit to validate a question, you know, based on a Reddit response, which may seem, which seems a little crazy, but the big thing is, that it's, it's what they learn, what they know at that point. It's not what they're going out and finding.

So if they don't know it, if they haven't learned it, so if your content is not on your site in a way that the AI can see it, And I use this analogy, I say, think about you're building a map for a city and there are no, the AI can't see any of the buildings. Because the map hasn't been really created yet. So those big buildings that are familiar to everyone, the AI can't see them. They don't have a schema, something that labels it past being an organization.

That's just what type of organization it is. They don't have the subject matter experts, which build incredible trust in their product. They're not connected to the product. The AI can't see that.

so because the AI, remember, is synthesizing the answer. Google is giving us the answer. 10 links, great. You can go through them, find what you want.

It's somewhere in those 10 blue links. Now AI is synthesizing and saying, this is the answer. If you're lucky, you're cited as well. If you've done your work, you're mentioned and you're cited, your URL is there.

And if it's there, you're going to see the people that end up going to your site. They're purchasing almost 20%. They're, they're pre-qualified. The ones that are, if you're mentioned correctly, if you're cited correctly, your URL is there.

And that often is because you've structured your site in a way that The URL was visible. Everything was visible to AI. um And that's the big shift. And that's why the invisible, because the AI, I'm sorry, know, the AI, you know, we forget the AI can't read the way we read.

Everything is numbers that is tokenized. So if you don't have something that says to them, hey, over here, our site and it's written in a language that they understand, they're going to miss it. They're not going to put that together. And I want to be clear for our listeners.

We're not just talking about the corpus of knowledge that's been ingested by the AIs as part of their training and as part of their foundation. We're also talking about the impact of what we would call deep research tools as well when they're scanning the sites, correct? Yeah, and actually the other piece of that is just your social profile, your listing on LinkedIn, all of it needs to be cohesive because the AI does get confused easily, also loses trust in you easily. um Target, uh the store is a great example.

I believe they'd fix this, but in December, 2025, Target was still showing up with a question of whether you want to target the store because of the entities being confused about the entities. Were you looking for a target, a physical target, or are you looking for target the store? And Bronco is the same thing. The Ford Bronco, we had to do a lot of work because of the horse and the team, which had a lot a lot more going for it than the Ford Bronco at first.

We had to really build up the right content because that sports content is naturally structured. It's within structured, it's stats. It's things that the AI can read easily. So it's very structured.

So we had to create Bronco information that was structured and separated the car from the team. So in our prep call, you and I had talked a little bit about this and I think you threw out the idea that about 20 % of legacy SEO work is what's visible to the AI engines and about 80 % is not. Is that, is about correct? and SEO will always be there.

And that's the structure that's keeping you in the game at this point, is that 20 % where you correctly build a foundation for your brand. And that's what it saw in its initial learning. That's what it's maintaining. But we do a lot with large-engaged model training, where you're creating posts, you're writing articles, you're on LinkedIn.

high trust environment for the AI. And you're using that to reinforce your trust, but also certain products, certain things that are happening at the company. can, as long as you're cohesive in your message, you can use these other really trusted structured spaces, Wikipedia being the most trusted, to train the LLMs. Let's let's make this real.

uh I know that you had ran an audit on. think it was roughly a $3 billion industrial software company, and I think you had said to me they were basically invisible to AI search compared to their direct competitors. So walk us through what is an audit like that actually surface? And I think for our C suite listeners, what does quote invisible look like when you put it up on a screen for an exact?

Yeah, so in that case, um the data is available. So going into the data and looking at the number of prompts that are being returned with the company mentioned, the company cited, and then breaking those down further to look at the product. In this case, it was a SaaS product with multiple products. So I think we got past.

So you start to look at all the different products. are they being mentioned? And in this case, um and the AI will also tell you the universe. um So if uh CrowdStrike or Palo Alto are number one, let's say, in cybersecurity, the universe is based off of the amount of uh prompts that could show up for them.

And the data, SimRush, any of them, data services create that universe. And then we run it against different, we have our own and often even use Claude just to go through the data and say, this universe really correct? So you get back a universe of cybersecurity questions people are asking about that software. And that universe could be in a month, a hundred million people, a hundred million prompts, you know, or questions.

that were asked. And the data will then tell you, how many were, if they're the top, CrowdStrike, Palo Alto were mentioned here. And then you look down and then there's another company, not in that specific space, but in another area of SAS that I couldn't think of the top people in. they should be the top, actually.

They were in 1,800. And only about 900 of those, they were cited. Your cited numbers should actually be a little larger. This dimension should be duplicative because it's putting them into buckets right now.

It'll be personalized and separated. It'll have millions of prompts to look at in the future. But right now, it's at least putting them into kind of categories. So the Other brands in that same space were seeing 225,000, I think was the top, for mentions uh in the space and right around there, about 185 in terms of citations.

And they're at 1,800. So that's going invisible. You're getting that 20 % basically of the Google search, the original SEO. uh IBM is a great example.

If you search what is blockchain, that's my favorite example, on your computer, you'll get a fully optimized IBM. They're usually the first result. And the page is exactly the way the AI needs to see it. The top, there is a question.

What is blockchain? It's matching that question. But then it's going, it gives a summary of what's on the page for the AI. And that's, you see it now on top of a lot of pages, under a hundred words, and it's paragraph at the beginning of the page.

It's a summary of what's on the page. now, usually down the left-hand column, there are the H2 headings, the kind of bold headings that down the article, they're H2 headings. so that the AI knows which in terms of parsing where things are. If someone has a blockchain or a question, that's not what is blockchain.

It's a little deeper. It knows from that little directory that IBM has put into the page. So there's now things on the page that are just for machines and things for humans. And that's the way your pages need to be designed.

I know that this is a little uh granular, maybe for some members of our audience, but when you talk about H2 headings, we're talking about H2 headings uh represented in an HTML format because I know there's also a lot of folks out there that are talking about attaching markdown files to websites as well. Yeah, so you mentioned I have two patents. Whenever I list my patents, I have to provide a link now to the listing in the US Patent Office that mentions my patents so that the AI knows they are real and will surface them.

had uh Jim Farley uh reference the work I was doing in a thank you. I click the thank you, I put it into one of my pages and I have recognition from uh Ford CEO and it links you to the video. A human may never see that or not even know it's there. Well, and I think that's the thing that interests me the most is this evolution for chief marketing officers and others about designing content that's made for human buyers.

But now we've got to design content that's made to surface oh our company and our products and our services to the level that then we get the attention of the human buyers uh as opposed to designing with the human first kind of mindset. Yeah, they're the new gatekeepers and that's, you know, if we were making a side list of the changes, I should have said that's right, you know, the big change. And so they're the gatekeepers of what the humans will see because there's, that response is synthesized.

So it's no longer 10 links. The human is not in control of selecting what they want to see. If the AI can see it, understand it, recognizes it's true and trustworthy, they then parse it and the human sees it. And that happens in that millisecond.

And we did a test on, I've mentioned this to you before too, on the Atlas browser, which is an open eyes browser. And we were looking for headphones and I, in the prompt, I mentioned, I'm looking in the test, I'm looking for headphones for my son that's in college. And it was like Samsung was not mentioned. Bose was not mentioned.

This small brand that had on Reddit a sub Reddit about college students saying how dependable this headphone was. And it had 0.6 % market share. It was returned as the suggestion for that I should buy.

That is wild. So if I'm listening, maybe I don't know, I'm a CMO, I'm a CFO, CEO, EIEIO, whatever, I'm listening to this episode and I'm wondering, do I have the same type of problem or maybe the same type of exposure with my company and my branding that we've just been articulating here? What's the cheapest signal that they can check tomorrow? I know you mentioned SEM, Rush and some others, but what's the least expensive way that I can figure out if I actually have this type of problem?

Cause you probably do, candidly. Yeah, so I'm involved with a, I'm going to mention another company that uh I just started working with. think they're doing something really um incredible. with uh basically with this entity layer.

And it really looks at whether what it's focusing on the AI agent, because I mentioned that's the real, the big difference that's on the horizon is the agent can purchase. So. That recommendation and college students are now asking AI assistants for their shortlist and it can get the application. know, there's so much so uh agent base AGNT doesn't Yeah, agent base agent base uh and.

Yeah, AGNT BASA. They have a free evaluation on their site and. No, a lot of them have you sign up to sign up for different things. They have a free, you just put your URL in, know, give you back how you look, what AI can see.

My personal site came back with an 82, which I kind of was, I was expecting, but I thought it'd a little higher. I had two scheme. I had two different things on there that were, uh, talking to each other that were messing things up. I had put something in when I shifted over to eo4.

ai and I didn't take out some of the forward information. So it picked it right up though. And basically I was confusing the AI is what it was telling me. I needed to pull that out right away because luckily I had enough areas where I was reinforcing that, but my personal site, I wasn't.

Let's go a little deeper on something you touched on a minute ago in terms of us moving to an agent purchased uh economy or an agent purchased model. Because when I surface this with a number of my clients, I either get looks of absolute horror or I get looks like I've got three heads, uh especially from the folks that have got thousands of SKUs and are moving multiple tons of product. I think about some of the distributors I work with, some of the folks that are in manufacturing and they are absolutely either terrified, appalled or disbelieving that we can move into a future where we've got AI agents ah in a B2B setting that are purchasing things on behalf of companies.

um And so I want to talk about that a little bit more because I do think that is the trend, the trend line here. And I think that's where we're going. And I'd love your perspective on this. So know I'm kind of leaving this wide open for you to step through.

Yeah. so I'll give you a consumer example really quick. One of the guys I worked with at Ford uh called me as this was happening. His assistant um was booking a night for her friends through ChatGPT and I guess they have the paid package.

So she was using the work, uh ChatGPT, to book a night out with her friends, gave the credit card, And it was making the reservations for her. It was booking the different places, saving everything. And then at the end, sent out an invitation to all of her friends. And she was done.

you know, it had done everything. The issue with the companies with a lot of SKUs and these large, um you know, whether it's Walmart or even, you know, a B2B um distributor. Um, these are now faceless pages because they don't have any real structure there. Uh, it's a product it's.

And it's in this, and it genuinely is invisible at that point. It's just sitting out there. And that's why I like this agent base. Their, their focus is on really diving down deep into that and the canonical, uh, canonical layer.

Um, And that level of structure, because the agent, you've got, you know, at least with where we are with AI search right now that we all use, it's not agentic. You know, there's a human in the middle where the human often, but there's human in the middle with the agentic piece. Now the AI agent is, has to be able to see it, trust it, understand it. And, um, Anthropic just did a test.

That was really scary. They did different types of models that had different levels. They were better models, basically. They're not releasing what the difference was with the models.

But they put them in an environment where they were able to purchase. And the better models were able to get better prices. do you know whether I don't think they negotiated, they just searched differently. They were able to find em and get better pricing.

And it was really scary because the models are advancing. so on the B2B side, even back to your original question, the big problem is if you're a SaaS company and someone has been tasked with recommending three products, they're going to ask AI uh They're going ask, hey, find me these three products, give me a recommendation on each one, why they're different. If that information is not in a way that the AI agent can see it, if it's not structured, and basically, um you have to build a knowledge base um within your site.

And it's something that's done all electronically, but... um It's done on a triple. if it's a SaaS product, insurance product, it does whatever it protects. And then who the clients are would be a triple on that.

the agentic AI needs to understand that the second they get to that page. What's the product? Who's it for? Then it will look at trust.

So if you have subject matter experts, their credibility needs to be then connected. And it's all through uh meta content. It's on the back end a lot. I'm sorry, it's connected in both places.

um But the meta content is that agenda is reading, is trusting, and you're connecting to if there's a Wikipedia page that is also connected in there. I'd recommend that I suggest everyone get a wiki data. don't think you can qualify for Wikipedia, set up your Wiki data page. It has a great level of trust and it's just a process you go through.

It's kind of the pre-process for Wikipedia. And those really trusted, I have an IMDB page because of some TV I did years ago, but it's a very trusted database. So I make sure that my IMDB page is correct. and that it's linked to everything because that's just one more layer that adds to my trust.

And where can we go to see your episode of Days of Our Lives? It was actually a hard call, which isn't even around. yeah, yeah, I remember hardball. OK, alright, uh I want to go a little uh deeper, but in a slightly different way on what you're talking about here.

um We're throwing around a number of different things. We're talking about schema and schema engines. We're talking about entity first talking about knowledge bases, and I think you've done a very good job of leading us through what this looks like. What I'm wondering here, because I know what some of our listeners are wondering, and that's how much of this is a tech project?

versus a content project versus a knowledge management project and who actually owns it when we're talking about doing this because we're at a really interesting intersection here I think. Yeah, it is, do think it is marketing has to be involved because you could take, I not to mention it, but uh a large data provider for businesses that has acquired uh a company for a specialization within their offering. And they've bought a company that's only important to people you B2B in that area.

You've now integrated that company and you've lost the name. You've now lost that to AI as well. So that those are marketing think, you know, thoughts that a marketing person would have, not necessarily that a data person would have. So laying that out, I use that as an example, kind of to the importance you bought it for its brand.

If I don't include that brand on in a way that AI can, know, everybody in B2B may know that you now own that and that's great. And that we, there was a good job with that. It's not truly B2B, but Mac Cosmetics was a great example. We worked on a project for them recently.

They have a big sponsorship with an artist and they did a great job of letting AI know that they have this partnership with the artist, but there's 15,000 people a day. searching for what makeup that artist is wearing, specific makeup. And they didn't think about that at all. They didn't put those SKUs into the page.

They didn't structure them. They didn't add that information in. They just said, we have a partnership with this artist. And so those 15,000 people are not getting oh a referral to Max Content to their...

So they're not paying for that sponsorship, know, that partnership. Geez. That's a fascinating example. And it reminds me of what you and I talked about before this uh podcast episode in terms of your mindset.

And I think you described it as a mechanics mindset. And I'm wondering, where do most companies misdiagnose what's broken here and start fixing the wrong thing if they don't have the right mindset? Yeah, and my dad was a mechanic and he, know, I guess, thankfully, we, had to take everything apart, put it back together. And he just really pressed on me.

Only way to learn anything was to be able to take it apart, put it back together. And you're right. I mean, that, you know, the CMO, you know, the CMO, somebody in there needs to have that strategy, needs to understand that these things need to be connected. And even on our projects, I'll run the schema through Claude or Chachi BT and schema, schema.

org is kind of the directory just for folks that it's basically there's 900 schema that identify basically two AI what is on your site, you know, what is in In your website in your products and you can get pretty granular with it unfortunately, there's about 10 to 12 that are used and You need to be more grand companies need to be more granular today, so I'll run I'm not reading these hundreds of pages You're not? So I will run them through and I'll just say, listen, did they mention this, this and this?

Is this correctly in there before this is put into the page? Which is the reason why my page is confused. So anxiety. I want to make sure that uh our listeners understood what you just said and I want to make.

want to confirm my understanding as well. So schema.org has got a an agreed upon for lack of a better term uh listing of 900 plus different terms and schemas that are used. Yeah, they were created by Google, Yahoo, Yandex.

And similar to how model content protocol was created by Anthropic and then donated to the Linux fund. It was created as a free resource that can be used. So now it's really become the points on the map for the AI. you need to, people will have organization, that's correct, okay?

And this goes back to that 20%. The organization is correct, maybe it lists someone's title correctly, it's 10 to 12 kind of general schema tags. Today, the AI needs you to go much deeper to understand. I appreciate you walking us through that because for those of us that aren't immersed in the world of generative engine optimization or even have the name or the credential CMO after our name, I think that that probably is news to a lot of folks, especially in mid-market land that schema.

org exists and that you can go and actually kind of use and leverage that information out there as you go through and redesign your site for the AI engines. So thank you. you can check, you requested earlier, can also check certain things there as well. And it's all free.

um So. Good. All right. Let's shift gears here a little bit.

uh During our prep call, you said something that really struck me and I want to underline it for the audience. You and I'll hold you to this here, right? You said to me that you think there's about a two year window where mid market companies have a real shot at becoming the source that AI is going to cite in their category. And after that two year period, it's going to be automated and the leaders are going to be locked in and they're going to have first mover advantage essentially.

So walk. Walk us through this thesis. I want to understand why two years and why is mid market actually advantaged at this moment and not disadvantaged. Great.

And I have sitting here, um funny, I didn't know you were going ask that, I did. So A16Z, which, um and Dr. Jason Hurwitz, which runs almost 20 % of our VC funding in this country, very committed to generative engine optimization. And their definition is, and that means optimizing what the model chooses to reference, not just whether or where you appear in traditional search.

That shift is revamping how we define and measure brand visibility and performance. And then they go on to say that it's going to be the system of interacting with LLMs. And they've had updated articles stating that it's also when you do LLM advertising that it will be a part of that because it's still the LLM delivering the advertising. So it needs to trust you.

So you can pay. and just be listed, or you can be really integrated in terms of advertising if the model trusts you. And the two years, and things are moving. I don't think anyone, I didn't expect Google to launch their AI agents as fast as they did.

That was probably six, seven months faster than I think anyone expected. Two years has been this kind of window. um There are a number of... uh answer engine optimization measurement companies.

uh Profound is one of the biggest or the biggest. And they are measuring how you're showing up for answers and giving you recommendations of what you can do to show up. And I've had this discussion with a few uh VCs and a folks in the space, how much that's setting us back. Because if you're just focused on the answer, you're not taking the time to do the rest of the things you need.

But they receive the funding first and they're really out there. But that creates an opportunity for, all the big, I'm shocked at how many big companies are using these measuring tools. And it creates a real opportunity for the mid-mark companies to go in. And if you can um educate AI today and the agent says they're starting out, it's Just think of, mean, they will understand who you are.

They will trust you and AI compounds. So it only reinforces over time. if you're, this, um like that, if that headset manufacturer that I mentioned in the Atlas search would take the time now to really build out their brand for AI and their offering and the company, they could stay ahead. for a segment for a long time.

um So if you're able to get in there now, identify your company, do the structure, do what you need to, you'll see it quickly. And um it's in the data. You'll see what prompts you're showing up for, what the AI is giving you preference for um in terms of it has an understanding of your company. And when certain things are asked, it will give you as the answer.

And where that starts to get even scarier is agentic AI. And there's a whole nother piece of that, which is intent. So we're going to start to get, we're already starting to getting personalized answers based on Claude Chachapie's understanding of you is skewing how it's going to answer your question. So if you're a mid-market company that can be more agile, think about your personas and who your audience is, and really develop your human and machine information around that, you're able to speak to that personalization piece, the intent, and get it really grounded into the memory of the AI.

And then you'll need to continue to continue to post that reinforces that's the sort of LLM training that we do. You're doing things that continue to reinforce this is your position or if you need to change it. And some of the interesting things that are happening on the entity and is people m and agent-based doing this where they're, was fascinated by this. They're creating basically a locker, which is gonna kill me for describing it that way, but.

It felt like a locker to me where basically it's an information locker. And if something is said incorrectly about a company, they're able to point the eye through the, um, and through some of the structure on the site to correct the answer, which I thought was really devious on their part, but that's the kind of thing I mean, and you know, that can all go bad too. That has, um, negative connotations as well. which is more of a reason to have these things in place.

So we're, we're talking here. It's mid 2026 and that two year window that we're talking about. So depending on when you're listening to this episode, it's a sometime between now and mid 2028. I want to make sure that I'm clear and our listeners are clear.

You gave one or two examples there, but you talked about compounding and you, uh, I want to talk and just get a little bit more context around. What are the primary compounding mechanisms? Is it citations? Is it training data?

Is it link patterns? Is it the wiki? Is it all of the above? What's the primary compounding mechanism?

Sure, we um did a large project that uh was for the Great Salt Lake in Utah. 25 different organizations, universities, so much data, so much information that we had to make sure the Utah legislators only meet for six weeks. In those six weeks, they make all the decisions for the year. So we had to reduce misinformation, make sure that the their model, you know, if they were using private models and the public models were all, you know, reduced misinformation and that we were identifying the issues related to the Great Salt Lake and the timing and everything correctly.

um And in terms of compact, so, and one thing we also did for the humans is we noticed they weren't as interested in saving the Great Salt Lake as we thought. And we could see that in the data. There wasn't a lot enough search for it for the water level for some of these things that are really important so we created structured content so blog articles pages on the sites that were structured to do with skiing to do a snowpack weather um Things that if you live in Utah you're interested in and we made sure that We made that connection between the Great Salt Lakes level and the snowpack and the weather and these different things.

And what I've been really surprised at is those prompts are still staying. We didn't lose those and the weather comes out every day. um We didn't end, there's articles about the ski pack and all these things, but they're not structuring them so we're... It's staying and it's continuing to gain momentum.

Now in order to keep it, they're going to need to do work. But that's an example of the compounding happening on its own. We only did those for the six weeks and we put them in place so that during that time, if people would see this information, if they were interested in what the, know, things with the weather, things, snowpack things and It's still showing up for now. It could still show up if everyone's committed to keeping that part of their universe.

um It could basically, it has a lead. It's on them. So. Great example.

And uh on a personal note, as someone who enjoys skiing in Utah, number one, thank you for that work. And number two, as somebody who's actually headed to Utah for business next week and will be down in the plains, uh anything that you can do to eliminate the arsenic and the other toxic dust that's getting uh ripped up into the air as a result of the dehydration of the lake, much appreciated. So thank you. We um increased public engagement by 400 percent, but what the big thing, they got the legislature, and this was a daily battle of keeping that information tight and um reducing the misinformation.

Probably in a year, they'll have a team work on the misinformation, but we were able to get ahead of that. um $100 million from the state legislators and a billion from federal to really start to, they know what the issues are. I mean, we'll separate, talk about that, but they know what the issues are. And this gives them, can, the 500 million that, you know, is said needed for the farmers to reduce their water usage.

There's the, you know, there's, it's enough money to get something done. Perfect. All right. Um, we're in the back half of this episode.

In fact, we're probably in the, the final quarter. So let's switch gears here again and let's talk about what not to do. I'm interested in the common, or maybe the most common mistakes that you're seeing mid market organizations or just organizations in general make and do right now when they decide to do something about AI search. Um, cause I think there's, um, a lot of actions that are being taken out there right now.

that look like progress but are actually compounding the problem and making it worse. So what are some of the common mistakes you're seeing in this area? Yeah, one of the things is um they're forgetting that AI, um so social media is not structured per se, your name is structured. um If you can add structure, especially by putting your URL and putting these things in, just keeping things on your website and making changes to your site is you need to reinforce that with a social media post that has a link to that page or, um, you know, making sure everything is cohesive and, um, that you're, um, you're breaking these things down.

You're not just saying this is the product. You're actually going into the details because this is the product. It already knows, you know, 20%. Um, and I've seen some companies really creating product pages and create structure within those pages for the same information that was already structured.

um But going into the detail, the trust piece is much bigger than I think people realize. If you have subject matter experts, if you are a subject matter expert, making sure you're connected to the product, to the page, so that you're credible for the AI. And it sounds so crazy that you need to be credible for the AI. That's what the model is looking for.

it's, and that's that validation, you know, needs to be able to, and it needs to be able to do all these things easily. So internal links, very simple things, you know, you can do to um help it validate in that millisecond that you have. That's a, that's very helpful. And I appreciate that advice.

And I think whenever we have a new technology come along, there's always a gold rush and there's always a lot of folks out there. Maybe they don't have ill intent, but there's a lot of folks that take advantage of what we don't know. And so I am personally seeing a ton, especially if you're on LinkedIn or anything like that of vendor pitches, and product pitches and things that Candidly, I think a lot of companies should be skeptical of right now. So I'm interested in your perspective without naming the guilty, ah is there a vendor type of pitch or is there anything out there that you're seeing that folks should be especially skeptical of right now in this space when we're talking about uh generative engine optimization and uh tailoring for the AI search tools?

Yeah, uh you want to see really concrete suggestions. You'll see a lot of folks that will do an uh AI check of your site, and they'll come back with a score. And they're going to tell you that this content, these very minor things need to be changed. Unfortunately, I'd be skeptical of that.

I've seen a lot of that, where they're getting a contract to come in and fix things. but they're not fixing the real structure of the site. And that's part of that answer engine, because you can manipulate an answer today. It's the generative part.

You need that structure. you could get a million. Your site could get a 54%, a 70%, a 90%, a 60%. And they all have very similar suggestions.

You want the one that is looking at the structure of it as well as the content. um I run a few test sites. uh we run, you mentioned these other products. We run one that instead of uh looking at Adobe or Mid Journey, it looks at the ones under that.

And we write articles about those companies to see how that search, it's just, but it's, can, it's, there's so much to learn because people that are coming in for those are looking to buy, which is crazy, you know, from these companies and they're buying a subscription for $20 because they don't have the hundred dollars, but really the 20 isn't worth anything. um So we're just gathering knowledge in from these sites. And one of them just does that looking at that. consumer behavior.

I appreciate that. Thank you. And again, names have been changed to protect the guilty. You had talked earlier, and I wanted to circle back to this very briefly.

You talked earlier about Google's rollout of agentic AI, and we talked a lot about optimizing sites for uh the AI buyer versus the human buyer. And one thing I neglected to ask you, but I really should have, is I feel like there's going to be a bridge period where we're optimizing both for AI as well as humans. So think about like an AI co-pilot and a human researcher working together before we get to that full AI buying uh paradigm or model, if you will. And I'm wondering, does anything that we're talking about here look different or change as we're in that bridge period before we get to true, fully agentic AI purchasing?

Yeah, I mean, a lot of what I'm talking about, unfortunately, is the bridge period and what you need in place now. um And those add ons that get you to the agentic, but you need the bridge in there first. um And it's really important um that you're always, um even, even just with AI in general, in my opinion, you need agency. Like, but I don't ever ask the AI to create just create this, I always give very detailed instructions or something, and you need to think about your site in the same way.

You want the human still to be able to understand it and have that level, and that only helps the AI in terms of it builds your trust. But I don't think, know, well, maybe, but in some areas, it will be full-ordented, but yeah, I think we're still need to keep that human in the loop. Yeah, no, I would. I would agree with that.

I I just I see things moving so quickly and I none of us know what the future holds and we don't quite know what the timelines are, but I feel like that's the direction we're going, especially in certain areas. There's, and so many companies are offering AI agents. That's a whole nother area. Um, I mean, Google has AI agents, chat, TPT is AI agents and now ad agencies, everyone is launched these AI agents and they have their, their function and you have to be careful of that as well.

Well, you you just said something that triggered a memory. was with a uh CTO that I work with. I work with a number of them for mid-market companies and they mentioned to me that their marketing department, the leadership in the company recently decided to reduce them by a factor of four in terms of headcount. And I wonder how short sighted that is.

given the changes that we're talking about and the complete overhaul that the schema and the underlying structure with content needs on the websites. And I think it's very seductive to think, we can just replace the marketing team with AI and we'll crank out all this stuff per our brand book, but it's not that game anymore. No. a bunch of studies that, in fact, the content was the bellwether on that.

And Google uh had their uh December update where they reduced AI content on sites that uh they have their EAT, their quality index. And they reduced a whole bunch of sites lost pages and pages where they had AI's slop. know, someone convinced them to put AI written content exclusively onto these pages without human review and Google, which is also interesting. The AI is taking out the AI, but it needs to recommend quality.

Yeah, well and as somebody that used to love the dash and has now had to surrender that uh I feel that so uh Alright I uh I like to to end our programs with a couple different types of questions and one of the questions that I always like is for something actionable if I'm a leader of an organization and I'm listening to this episode We haven't done anything yet on generative engine optimization But maybe I'm listening to this episode on my drive in on Monday. Does anybody drive it anymore?

I don't know. But I'm walking into my leadership meeting. What's the one thing that I should be thinking about putting on the agenda this week and not something that's going to be a six or a 10 month transformation, but just a move that we can make. Yeah, so I mentioned um and I don't have any ownership in this company at all.

I just think it's a great technology, but it's AGNTBASE. You can go on there for free. can get the look at the structure of your site and tell you about that. And then they have a $30 fix that they offer, um which is worth every dollar, you know, just if you get a page fixed, and then you know what how the rest of your pages should be.

But I think that's an easy one. They have a free and starts at $30. So can't really lose there. No, that is for sure.

That is a wonderful price point for that. Something else I'm interested in. What are you most excited about now, Alan, because you see an awful lot. uh You're in the trenches with the companies.

You're building the tech of the future today. What are you most excited about right now? and this, what I'm most excited and I think this as a small business person myself, and as someone who's been a founder, I am excited. It's counterintuitive in some ways, but there are some great new technologies that will create your storefront, your, put your product in place and all these things that are tedious and you know, as best for repeatable tasks.

So. There are some cool new products that uh do that for you, which I think is really neat. Very good. All right.

Alan, an hour always goes pretty quickly here. We're just about coming up on it. I want to make sure I ask if any of our listeners want to reach out to you, learn more about what you're doing. I think you may have a book out there in the wings somewhere.

uh How can they find out more about you and get in touch? Sure. So my website, it's EO4.ai.

that stands for Entity Optimization for AI, uh which in my opinion is the most important thing that you can do is making sure your entities are clear to the AI. And that's covered on the site. there's a blog on there with some tips. And my LinkedIn page is also chock full of uh recommendations and tips that you can use.

definitely um you can email me. You can contact me through the site. My email is myname, uh alan at rambam.com.

And Alan is A-L-A-N. Alan at e04.ai works too. Wonderful.

Alan, ah I would imagine that we're going to have you back on the show because of the pace of change and how fast this world is changing. And I really appreciate you sharing your hard won. And I do mean that it is hard won knowledge with our listeners today. So thank you very much.

thank you so much. It was really great. can't believe it's over. Yeah.

Absolutely. Alan Rambam, everybody. And uh for those of you out in listener land, uh thank you for listening once again to AI for the C-suite. If this episode was useful, uh check us out and subscribe wherever you get your podcasts.

Follow us on LinkedIn, visit AI for the C-suite.com. And until next time, keep your algorithms running, your leadership evolving and your AI in check. Take care, everybody.