What the top SEO and GEO specialists actually say about AEO and GEO

The specialists are split, but not on everything. On the naming and framing, they diverge sharply: Mike King argues AI search is a genuinely new discipline he calls relevance engineering, while Rand Fishkin and Lily Ray see much of GEO as rebranded SEO. On the fundamentals, they largely agree: E-E-A-T, entity authority, citations, and clear direct answers still decide who gets cited.

This piece reads the field fairly. It presents both camps, quotes each voice from a source you can check, and separates verified statements from paraphrased positions. The goal is not to declare a winner. It is to give you an accurate map of a debate that is still live, so you can decide what to actually do about it. For the terminology itself, see SEO vs AEO vs GEO vs AIO vs LLMO.

The two-line summary

• The "it is a new discipline" camp is anchored by Mike King, who says the SEO frame is too limiting for how generative systems transform inputs into answers.
• The "it is still SEO" camp is anchored by Rand Fishkin, Lily Ray (in part), Britney Muller, and others, who see most GEO advice as long-standing SEO under a new label.

Both camps agree the underlying work still rewards authority, entities, structured answers, and being safe to cite.

Rand Fishkin (SparkToro): it is still SEO

Fishkin is the most quotable skeptic of the new-acronym wave. His argument is not that AI search does not matter. It is that the discipline that serves it already exists.

"AIO, AEO, GEO, LLMEO, and the other 15 I saw on LinkedIn today are not the way."

"Search Everywhere Optimization is good enough. In a few years, I doubt you'll even need to explain that you help brands influence people outside of just Google rankings."

"I'd bet if you've done SEO for the last few years, you're already doing Search Everywhere Optimization."

His position: the proliferation of acronyms is noise, and "Search Everywhere Optimization" (a framing he credits to Ashley Liddell) captures the real job. Source: It's Still SEO: Search Everywhere Optimization (May 29, 2025).

Two related Fishkin/SparkToro positions matter here. First, on measurement: SparkToro and Datos have paraphrased that AI answers are so non-deterministic that there is less than a 1 in 100 chance that any AI tool asked the same question 100 times returns the same brand list in any two responses (Search Engine Land). Second, on traffic: Fishkin's long-running zero-click thesis holds that most Google searches end without a click, so brand visibility can be valuable even when it sends little or no referral traffic (In a Zero-Click World, Traffic Is a Terrible Goal).

Mike King (iPullRank): relevance engineering

King is the sharpest voice for the "this is genuinely new" position. Named Search Engine Land's 2025 AI Search Marketer of the Year, he argues the SEO label understates what has changed.

"The reinvention I'm calling is relevance engineering, because the SEO frame is so limiting."

"You put something into Google before, it comes out the same on the other side. Now it is enhanced. It is chopped and screwed and then it's served up to the user."

"The big difference is that what they do with our inputs is dramatically different... the way we get that isn't through optimization. It's through engineering."

His position: generative systems transform your content before serving it, which makes the work more technical than classic ranking optimization and, in his view, deserves its own name. Source: Mike King SMX Advanced 2025 interview.

Lily Ray (Amsive): mostly SEO, but a real new system for competing and measuring

Ray occupies the middle, which makes her the most useful voice for readers who distrust both hype and dismissal. She is blunt about opportunists.

"many such GEO grifters were using this opportunity to simply repackage core SEO approaches using a different name."

"verbatim recommendations that SEO teams have been making to their clients for years."

But she does not conclude that nothing changed.

"Ultimately, AEO/GEO is not an overhaul or abandonment of SEO. Instead, it represents a new system for competing for, capturing, and measuring success across AI platforms."

And from her Affiliate Summit West 2026 keynote:

"Search has evolved into answer. We're no longer optimizing for 10 blue links. We're optimizing for AI-generated answers, agentic commerce, and brand visibility across large language models."

Her position: the tactics are largely familiar SEO, but the competitive and measurement surface is new. She prefers the term AEO, and she holds that E-E-A-T remains the foundation. Sources: A Reflection on SEO and AI Search and The State of AI and SEO in 2026 with Lily Ray.

Aleyda Solis (Orainti): frameworks over labels

Solis has focused less on the naming fight and more on giving practitioners a repeatable process. She built the free Learning AI Search roadmap (learningaisearch.com) organized around a three-layer framework: Presence, Readiness, and Business Impact.

Her paraphrased content principles are concrete: lead with a direct, concise answer sentence; include a key-takeaways summary; and keep a plain, factual, non-promotional tone. She also ran a survey of more than 200 senior SEOs and reported that the industry does not agree on a single term for this work, while the top tactics converge on schema and structured data plus digital PR and citations (Reddit, Wikipedia). Source: AI search optimization survey 2025.

Marie Haynes: making brands safe to cite

Haynes, author of SEO in the Gemini Era, was among the first to document Google's query fan-out behavior in March 2025. Her paraphrased focus is on trust signals and making a brand safe to cite, which reframes AI visibility as an authority and reliability problem rather than a keyword problem. Source: profile background. (Her specific AI Mode remarks have circulated second-hand, so they are represented here as paraphrase only.)

Britney Muller: non-deterministic machines, deterministic index

Muller's paraphrased argument is a useful technical grounding for the skeptic camp. She frames LLMs as non-deterministic probability machines, but notes that every URL an LLM surfaces still comes from a live search index. If retrieval depends on classic crawling and indexing, then AI visibility is, in large part, rediscovered SEO.

The neutral reporters and the pragmatists

Two more voices round out the picture without picking a side. Barry Schwartz (Search Engine Roundtable, marking 22 years in 2026) is the field's neutral daily chronicler of Google AI Mode changes; his value is documentation, not advocacy. Gianluca Fiorelli takes a pragmatic line, treating AI search as one channel inside a broader demand-generation and semantic-SEO strategy rather than a separate discipline. Dan Petrovic is a sharp skeptic of the original Princeton GEO paper's rigor; because his critique is reported second-hand, it is represented here as paraphrase.

The academic anchor: the Princeton GEO paper

The term GEO has a specific research origin. GEO: Generative Engine Optimization was published at KDD 2024 (ACM SIGKDD) and is available as arXiv:2311.09735. The authors are Pranjal Aggarwal (lead, IIT Delhi), Vishvak Murahari (Princeton), Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan (Princeton), and Ameet Deshpande (Princeton).

The paper introduced GEO-bench, a benchmark of 10,000 queries across 9 domains, and claimed that targeted methods, such as adding citations, quotations, and statistics or authoritative data, can boost visibility in generative answers by up to 40%. Two caveats belong in any fair reading: it was non-peer-reviewed at its first posting, and several practitioners dispute how well its results generalize beyond the benchmark. The paper is worth knowing because its recommended tactics, cite sources and add data, map almost exactly onto what the practitioner camp already endorses.

Where they disagree

Three genuine fault lines run through the field.

Is GEO a new discipline or rebranded SEO? King says new (relevance engineering). Fishkin, Ray (in part), Muller, and Petrovic lean toward rebranded or incremental. Ray's middle position, familiar tactics on a new competitive surface, is the one most people can live with.

Does AI search send meaningful referral traffic? Fishkin argues it is tiny today and that visibility, not clicks, is the point. Others are watching the trend line rather than declaring it settled.

Are AI-visibility metrics valid given non-determinism? Fishkin and Muller are the skeptics: if the same prompt returns different brand lists on repeat runs, single-snapshot metrics are shaky. This is a measurement design problem, not a reason to ignore the channel, but it should make you distrust any tool that reports a precise, stable "AI visibility score" from one query.

Where they agree

Strip away the naming fight and a consensus emerges. Across skeptics and enthusiasts alike, the same fundamentals keep coming up:

• E-E-A-T remains the foundation (Ray, Haynes).
• Entity authority matters: be a recognized, well-described entity, not just a page. See entity authority for AI engines.
• Structured, direct answers win: lead with the answer, then support it (Solis).
• Citations, quotations, and data measurably help (Princeton paper, Solis).
• Digital PR and third-party citations (Reddit, Wikipedia, credible mentions) build the off-site signal (Solis survey).
• Being crawlable and safe to cite is table stakes (Muller, Haynes).

That is a striking amount of agreement for a field that argues so loudly about what to call itself.

What I take from all this

My read, after weighing both sides fairly, is that the argument is mostly about framing, and the framing matters less than the field's volume suggests.

King is right that generative systems do something new: they transform your content before serving it, and that is a real technical shift worth respecting. Fishkin is right that most of the day-to-day work is continuous with good SEO. Ray's synthesis reconciles them: same fundamentals, new competitive and measurement surface. You do not have to pick a tribe to act correctly.

Practical takeaways:

• Do not wait for the naming war to end. Ship the work the fundamentals demand. It pays off regardless of which acronym wins.
• Treat E-E-A-T and entity authority as the base layer, not an afterthought. Every camp agrees on this.
• Structure content to be extracted: a direct answer up top, a key-takeaways summary, supporting data and citations below.
• Earn third-party citations, because AI systems lean on what others say about you, not just what you say about yourself.
• Distrust precise AI-visibility scores. Given non-determinism, treat them as directional trends across many runs, not exact readings. Measure share of voice over time, not a single snapshot.
• Stay crawlable and safe to cite. If a system cannot fetch you or trust you, none of the rest matters.

The honest conclusion is unglamorous: whether you call it AEO, GEO, relevance engineering, or Search Everywhere Optimization, the winners will be the sources that are authoritative, well-structured, frequently cited, and genuinely trustworthy. For how the discipline got here, see the evolution of search from SEO to GEO, and for the AEO-versus-GEO distinction specifically, see AEO vs GEO explained.