AEO, What You Need to Know, Part II - Four Columns Marketing

In Part I, we talked about what’s going on with SEO and AEO at a high level. In short—AEO is gradually taking over more and more of the role that “Organic Search” has typically held.

In this blog, we’ll be addressing what makes them different from a technical perspective.

It’s important to establish that (in my opinion and hindsight) this is always the direction that organic search was headed. To back that up, let’s talk about how SEO, generally, works:

SEO, Oversimplified

SEO has gone through several major evolutions since Google launched their search engine. In its infancy, it was possible to game SEO algorithms by “keyword stuffing.” You could throw a word on a page a couple dozen (or hundred) times and the crawler that set rankings would use basic frequency count to assess relevance. Those were simpler times—and the internet caught on. Keyword stuffing became standard practice, so Google made updates.

The purpose of this blog is not to provide a condensed history of SEO, so I’ll spare you that deep dive. If you want to know more, SEMRush, one of the leading SEO platforms in the world, publishes SEO contributors regularly.

What all of these changes amount to is important: trust.

Google’s algorithm for SEO uses a wide-ranging set of parameters to establish which pages are most relevant to a given search. This isn’t just about the words on your website. Google looks at elements like:

How fast does your site load?
Is your website optimized for mobile phones?
Do other websites link to your website?
Do you have the right technical implementation on your website?

Google breaks down how they rate/judge trust in search rankings through E-E-A-T:

Experience, Expertise, Authoritativeness, Trustworthiness

This acronym matters for everyone. It asks strategic questions that are answered by your implementation—or lack thereof. It looks at more than your website. It’s looking at your business directory listings, your reviews, how your allies reference you and much, much more.

Its more uptight cousin acronym is YMYL (Your Money, Your Life), which is an added layer of scrutiny for industries with the largest impact on searchers (Finance, Healthcare, Legal Advice, Safety). Take everything EEAT is doing at an ‘8’ and turn it up to ‘11.’

So this system originally based on keywords is now looking at… keywords plus everything else?

Yes. If you’ve ever written a page that you expected to rank and found yourself disappointed when it didn’t, it’s likely your problem exists in one of the other 100+ parameters SEO is using to judge your relevance.

Is this system inherently fair?

Yes and no. Yes, because by diversifying the data points that inform rankings, Google has been reasonably effective in limiting bad actors’ ability to game the system. No, because it makes it almost impossible for small businesses with poor marketing knowledge to show up without expert intervention.

Isn’t this supposed to be about the technical differences between AEO and SEO?

Yes, and we’re getting there—but I think it’s important to understand that the technical differences in AEO and SEO aren’t really about throwing the entire system out and starting over, it’s about the next evolution in an already complex and multi-faceted digital ecology.

One last one—and this is important: how do SEO/Keyword and AEO/Query searches differ?

“Best Running Shoes Waco” = Keyword Search
“Where is the best place to buy running shoes in Waco?” = Query Search

Not To Be Dramatic, But…

I’ve spent a fair amount of time on stage, both for music and as an actor. One of the most fundamental tenets of theatre improvisation is “yes, and…” For the uninitiated, this concept confirms context (Yes) and adds additional instructions for others on stage (And).

Before we jump into AEO—think about this concept in light of how AI always follows up with you. “Would you like me to draft that email for you now?” “Want me to outline that page so your copywriter can keep it going?” “Want me to check on availability for that product?”

I don’t think I’m exaggerating to say I see a parallel here.

SEO is most commonly a “one-off”—you search “Best Running Shoes Waco” and you might click on 1-3 results on the search engine results page (SERP) before you change some words in your search to get a closer answer.

AEO, by nature, is setting itself up for the next question.

SEO is “Yes, this is the strongest result for those words.”
AEO is “Yes, there are several strong results for that question, it looks like there are several stores in your area that have good options, would you like me to check which stores have sales happening right now?”

Here Comes The Payoff Pitch

You came here for the specifics, and it’s time to get into them. Let’s continue the running shoe example from the previous section and watch how this might play out in a purely hypothetical scenario.

Localization/NAP
One of the biggest surprises with AEO for me personally is that it seems to default to localization more than SEO does or ever did. On truly local searches like our test case, this can get pretty pointed.

Google’s SEO search and AI search actually return the same company—but in very different formats:

Screenshot of Google homepage search for "best running shoes waco."

Screenshot of Google homepage search in AI Mode for "best running shoes waco."

It probably doesn’t hurt that Waco Running Company is only a couple blocks from our office—but that’s not doing as much heavy lifting as a well-maintained Google Business Listing, strong review scores across multiple platforms and strong backlinks from multiple local sites.

That’s localization done well, and the answer AI gives is a LOT more robust than a first-position ranking.

While this example is (local bordering on hyper-local), localization isn’t just about “within 15 minutes of me” anymore. “What’s the best roofing company in my area” is likely to pull results from Waco/Temple/Killeen. “What’s the best BBQ within two hours” is likely going to stretch past College Station—but all of that hinges on businesses that have (1) rich, location-inclusive content on their websites and (2) a well-tended digital footprint that aligns critical information and third-party feedback like reviews across the full range of directories, digital assets and platforms. Easy, right?

Crawl to Run, Structure to Climb

There are two deeply technical considerations that AI elevates in the broader conversation around organic discoverability—how “crawlable” is your website and how well does your website tell those crawlers where to go and what they’re going to find there?

This isn’t new stuff—both have been part of the broader SEO conversation for years—but two things have changed: the stakes and the audience.

Let’s start with something basic.

Robots.txt

Every website has this file. It’s plain-text that sits at the root of your domain and tells your site’s automated visitors (yes, you have them) where they are and aren’t “allowed to go.” For the last couple decades, this was a two-party conversation between Google’s crawler bot and your site.

Enter a whole new class of visitors:

GPTBot (OpenAI)
ClaudeBot (Anthropic)
Google-Extended
PerplexityBot
And more!

These crawlers feed AI models that people are either (1) intentionally using or (2) encountering through changes in their search experience. Here’s something that might catch you off-guard: if you haven’t updated this file in the last two years, it could be a reason AI is having trouble finding your website.

If you’re a Four Columns client, your file has been updated to a configuration that prioritizes high-value crawlers, delays lower-value crawlers and outright blocks nefarious crawlers. Just part of the friendly service! The fix is pretty simple, and it’s something just about any AI model can help you with, just verify that you’re confident in which crawlers you should/should not be blocking.

Related, But Separate

You might have heard of llms.txt—I’ve certainly heard about it from every SEO influencer on Instagram—this is being pitched as a ‘robots.txt for AI’ solution. The idea is fine in theory, you get a clean markdown map that tells AI tools which of your pages are worth reading… but the honest truth is that the actual payoff isn’t proven. No major AI solution has committed to using these resources and while it’s not highly intensive to generate (particularly with AI assist), it’s also not a guarantee, and should be a red flag if it’s a huge selling point from an SEO/AEO company.

Schema

I, a noted overutilizer of analogies, have two more for you that might help you understand schema.

Have you seen The Matrix? Humans see the world one way, and the machines see it another? Not a terrible way to think about any of this, but particularly this concept, because schema is invisible to humans, but catnip for AI. We, simply put, see the world through different eyes.

Second analogy—think of a climbing wall:

Caucasian men and women in teens and 20s working together in practice sessions moving up vertical walls in sport climbing gym.

When you look at this wall, you see the whole thing. You see how tall it is, how the facade changes, and the overall path you’d take at a glance, because you, reader, are a brilliant human. This is how you take in the world.

The second you stop looking and start climbing, you cross into bot territory. You’re not thinking about the wall, you’re thinking about the next grip point and where you can go from there.

Crawlers of your website are climbers, not spectators. It doesn’t see your page, it doesn’t care about your colors and white space. It looks for structured hand-holds. Schema is structured data markup that lets you “bolt” these holds onto your wall. It’s a standardized vocab that’s invisible to your site visitors but explicitly labels pieces of your page: this is the business name, this is the physical address, this is a review rating, this is a product category, this is the price value, this is an answer to a commonly asked question, etc.

Yes, AI can infer a lot, but if Schema is catnip to AI, a lack thereof is handing AI a gummy—there may be some mild hallucinations. Schema eliminates the guesswork for AI and tells it exactly where it needs to go, because there is ultimately a difference between machine-visible (AI can see the content, but doesn’t have context) to machine-legible.

This has all been good practice all along. Crawlability? Structure? Those are age-old SEO concepts that pre-date AEO, but AI raises the stakes of getting this stuff wrong.

Hand in Hand

So much of marketing follows the “yes, and” principle I cited earlier, and this is no exception. Think of these two pieces this way:

Your “robots.txt” file is the doorway to your website—it says “Yes, [AI], come on in!”

Your schema markup is a guided tour with answers and explanations.

Lack the first one, AI never gets in the door.
Lack the second one, and you’re leaving it to wander around like a grade schooler separated on a field trip.

Together, these two pieces don’t just get you found, they make you a source that AI can confidently build an answer around, then keep building on the next question and the next.

Enough, Joe, Enough!

That’s enough for now, and serves as a strong technical foundation. In part III, we’ll talk about what all of this means for your actual marketing plan, where you should spend your time, what isn’t worth doing and how to think about AEO as a fluid strategy rather than a one-time checklist.

Or, like always, you could just call 4C and skip the homework.

Joseph “Joe” Gastler is the Chief Marketing Officer at Four Columns Marketing and an adjunct lecturer in marketing at Baylor University. He received his MBA from Baylor University with a concentration in marketing. He is certified in Google Search and has over ten years of experience managing SEO strategies, building websites and directing content mixes for hundreds of website projects.