AI Search Only Feels New If Your SEO Was Shallow: Stop Arguing GEO vs. AEO and Master the Fundamentals


Hi everyone, this is Neo.

The SEO world is arguing again: GEO, AEO, AAIO, AISO… every few months a new acronym pops up, and behind every acronym there’s someone selling courses, tools, and “if you don’t get on board you’re already too late” anxiety. I get similar questions in my WeChat Official Account inbox all the time: “Neo, the AI search era is here — should I throw out traditional SEO? Should I rush to learn GEO?”

Today I want to use a new article by Slobodan Manic on Search Engine Journal to sort out all this back-and-forth once and for all. His core point is simple, but it might make you uncomfortable:

AI search and Google search are two different races, but the training is nearly identical. The GEO/AEO naming debate is really an escape from a more awkward fact — you might have skipped the fundamentals back then.


1. What the original article says: two different races, nearly identical training

Slobodan Manic hosts the No Hacks Podcast and focuses on “machine-first” web optimization consulting. His analogy: the 100-meter and 60-meter sprints are two different events, possibly with different champions, but the training you do for both is nearly the same.

Google search and AI search work the same way. What wins in AI search is the same set of things that win when you do traditional SEO right:

  • Clarity: your page should make it obvious to people (and machines) what you sell
  • A healthy website: no messy technical problems
  • A healthy off-site profile: how others describe you, and where you show up

“Not one of these recommendations is written for AI,” he says. “They’re the unglamorous, plain-as-day advice that has been around for years — no courses to sell, no new terminology to invent, so most practitioners walked right past them chasing whatever tactic ranked highest at the time.”

There’s one key sentence I want you to remember:

When a model answers a question about your product, it’s doing what search engines have always done — finding the clearest, most consistent, most trustworthy narrative of “who you are and what you sell,” and repeating it. A clear website plus a consistent off-site profile makes that easy for Google and easy for language models. The mechanics changed; what gets rewarded didn’t.


2. He’s not the only one saying it: industry-wide consensus within days

Manic isn’t a lone voice. He names three industry figures in his article who reached the same conclusion from three different directions within days:

  • Jono Alderson (senior technical SEO consultant, on the No Hacks Podcast): “SEO vs. GEO is the wrong question — AI isn’t a new discipline.”
  • Mordy Oberstein (well-known SEO analyst, in conversation with Brent Csutoras): “SEO isn’t dead, strategy is dead.”
  • Ross Hudgens (founder of Siege Media): warns practitioners against labeling themselves “SEO/GEO writers.”

Three people, three angles, one conclusion. Manic puts it directly: when that kind of consensus appears, it usually means the naming debate has become an escape — an escape from a simpler, uglier conversation: you may simply never have mastered the fundamentals.

That stings a little, but I think it’s especially worth chewing on for independent store owners in China. There’s a very common pattern in our industry: the moment search traffic fluctuates, the first reaction isn’t to check content quality, site health, or brand consistency — it’s “did a new algorithm drop? Do I need to learn something new?” And then another round of harvesting by the new-concept machine.


3. What’s actually new?

Manic also admits one thing has genuinely changed: off-site authority matters more than ever — but note, not “backlinks” in the traditional sense.

He’s talking about two unglamorous things:

  1. Entity Optimization: models think in “entities.” They need to know who you are, what you’re known for, and which real people in your company are recognized as experts in the field.
  2. Consistency of your brand everywhere it appears: the model’s picture of you isn’t built from your homepage alone — it’s assembled from every place it can read: your website, your YouTube, your Reddit replies, industry media coverage, how others introduce you.

“If your SEO used to be content marketing — publishing shallow articles to rank temporarily — then AI search really is new to you, and it will stay new and hard,” he says. “Because the tactics that produce temporary rankings can’t produce a consistent entity a model can trust. On the flip side, if you’ve been doing the fundamentals all along, this isn’t new at all — you’ve been doing entity work and consistency work, just nobody called it that.”


4. Neo’s take: why I agree, plus one important addition

1. The “naming debate” is the easiest trap for Chinese independent store owners

I’ve seen too many sellers who, the moment they hear “GEO is the new SEO,” throw all their energy into “make ChatGPT recommend me” and drop traditional indexing, internal linking, and content structure entirely. The result? AI citations never took off, and Google traffic dropped too.

These two aren’t an either/or — they share the same foundation. Semrush’s 2026 research proves it: although close to 90% of pages ChatGPT cites rank outside the top 20 in traditional search (we’ll dig into that number in a later post), Semrush also found that LLMs cite pages ranked 1-5 more often than pages ranked 6-20 — meaning doing well in traditional rankings still significantly boosts your odds of being cited by AI. The foundation didn’t collapse; there’s just a new floor on top.

2. “Can the model actually use what you feed it?” is a badly underestimated question

At the end of his article, Manic gives one piece of practical advice that I think is worth more than any “GEO tutorial”:

Check whether what you’re feeding LLMs, Google AI Overviews, and AI Mode could cause them to hallucinate and pass wrong information to users. Then prompt questions about your brand and your products, see what the model knows and where it gets it from, and reverse-engineer it.

In plain English: stop guessing what the model thinks of you — go ask it directly. Ask ChatGPT, Perplexity, and Gemini about your brand name and core categories, and watch how they answer, who they cite, and whether the information is right. It’s the cheapest, most honest AI visibility audit there is, and no tool can replace it.

3. My full definition of “the fundamentals” (a practical checklist)

Since we’re talking fundamentals, let me be explicit. In my view, the fundamentals in the AI search era are these five layers:

Layer What it means How it maps to AI search
Content layer Every page has a clear, direct answer; titles and body copy don’t play games The model has to be able to extract “what you sell and why it’s good”
Technical layer Crawlable, renderable, structured data, speed If crawlers and models can’t read it, it doesn’t exist
Entity layer Brand name, founders, and core products tell a consistent story on-site and off-site Models assemble your profile by entity
Consistency layer Website, social media, marketplace store, and media coverage all describe you the same way Cross-platform consensus = credibility
Evidence layer Real customer reviews, third-party coverage, verifiable case studies Earned media carries the most weight

Across these five layers, not a single one is “new knowledge.” They’re all common sense Google has hinted at and stated outright for the past decade — people just don’t want to admit it, because admitting it means there’s no “secret AI search playbook” left to sell.


5. Before you switch lanes, ask yourself these five questions

At the end of the article, here’s a self-check list. If you’re agonizing over “should I go all-in on GEO,” answer these five first:

  1. Can AI crawlers read and render my site properly?
  2. Do my core pages have direct, clear, structured answers?
  3. Does my brand say the same thing on my website, YouTube, social media, and industry media?
  4. How long has it been since I last updated content — and was it a “date change” or did I actually add something?
  5. Last time I asked ChatGPT/Perplexity “what’s the best X brand,” did my brand show up?

If #5 comes up empty, don’t rush to blame AI search for “playing favorites” — go back and look at the first four. Chances are you buried your answers too deep, or you simply left no trace anywhere else.