The honest skeptic’s case against GEO.

Why this article exists

Most published GEO content is sold by people whose business depends on GEO being important. That is not a disqualifying conflict, but it is a real one. The strongest case against the discipline deserves a hearing on its own terms — not a strawman, not a polite acknowledgment before the inevitable rebuttal, but the actual argument as its proponents make it.

What follows is the most credible version of the skeptical position, drawn from named practitioners with deep SEO experience and data-backed reasons for their skepticism. Where the position is right, the article says so. Where the position over-reaches, the article says that too. The goal is the honest assessment a serious buyer should be able to make before committing budget.

Fishkin’s consistency finding

The most empirically devastating critique comes from Rand Fishkin’s SparkToro research published in 2025. The methodology was straightforward: ask the same brand-recommendation question of an AI engine, repeatedly, and measure how consistent the responses are.

The findings:

• Fewer than 1 in 100 prompt runs produce the same list of brands.
• Fewer than 1 in 1,000 produce the same list in the same order.
• The probability of identical lists in identical order is under 0.1% regardless of platform, vertical, or query phrasing.

The implication, in Fishkin’s framing: any tool that gives a “ranking position in AI” is full of baloney. AI engines are probability machines designed to generate unique answers every time. Treating them as deterministic ranking systems is, in the technical sense of the word, nonsensical.

This critique is correct. It does not make GEO irrelevant — visibility percentage measured across many prompt repetitions is a valid metric — but it does invalidate a specific class of GEO marketing claims that promise “rank #1 in ChatGPT” or “guaranteed Perplexity placement.” Buyers should treat any vendor making such claims as either uninformed or dishonest.

“Any tool that gives you a "ranking position in AI" is full of baloney. — Rand Fishkin, SparkToro”

Profound’s citation drift research

Profound — one of the largest GEO platforms by market presence — published research that arguably argues against the durability of its own category. The findings:

• 40–60% of domains cited in AI responses change completely within one month.
• 70–90% change over six months.

The interpretation: AI citation positions are radically more volatile than Google rankings ever were. A site that wins citation share this quarter has no strong reason to expect that share next quarter. The optimization investment may not compound the way SEO investment historically did.

This critique is partially correct. The citation drift is real and measurable. But the interpretation cuts both ways: the same volatility means challenger brands can capture share quickly, in ways that were structurally impossible against incumbent SEO authority moats. The volatility is bad for incumbents and good for challengers. Whether your organization sees it as a threat or an opportunity depends on which side of that line you sit on.

Why Lorelight shut down

The most damaging critique of the GEO industry came from inside the GEO industry. Ben Goodey built Lorelight, an AI visibility tracking tool, then shut it down in 2025. His public reasoning was unusually candid.

The conclusion that ended the company: “Everything companies needed for GEO was the same as what they already needed for good marketing.” Lorelight experienced significant churn because GEO “currently has no long-term strategy” — tactics were “hacks more than strategies.”

From Goodey’s exit interview with Section: “Companies mostly expected a magic bullet. They thought they would log in and it would say, ‘if you do this one thing, suddenly you will be number one in ChatGPT.’ Instead they got: ‘if you keep doing the hard work, then you will show up.’”

This is the strongest skeptic argument because it comes from someone who tried to build the product, learned what wasn’t working, and was honest about the result. The fair response: Goodey is right that GEO is not a magic bullet. He is also right that 80% of the work overlaps with foundational marketing. He is wrong, in the read of multiple operators who continued investing, that the remaining 20% doesn’t matter — that 20% is exactly where citation share is decided.

The snake-oil critique

Kai Spriestersbach, the AI researcher and SEO veteran widely cited in the European GEO conversation, made the bluntest version of the skeptic argument in Business Insider:

“Everyone is going crazy about becoming the next agency — all the side hustlers and snake oil sellers with their tools already on the train and riding the hype.”

The market reality validates the framing. 20+ GEO tools launched in 2025 alone. Several have raised at $100M+ valuations despite operating in a field where, as Spriestersbach notes, no one has proven, replicable methodologies. Lily Ray (Amsive) and Jeremy Moser (uSERP) have both publicly warned that “80% of GEO is good fundamental SEO” and that anyone claiming otherwise is selling snake oil.

Where the critique is right: the industry has attracted bad actors. Where the critique over-reaches: 80% overlap means 20% net-new discipline, and that 20% is where competitive positioning gets decided. The buyer’s test is not whether GEO has snake oil — every emerging discipline does — but whether the specific vendor in front of you publishes their methodology, cites primary research, and can show measurable lift attributable to their work.

Where Google’s own representatives draw the line

The most institutional version of the skeptic position comes from inside Google. The platform’s public-facing representatives have been remarkably consistent:

• John Mueller (Search Advocate, January 2026): “There is no such thing as GEO or AEO without doing SEO fundamentals.”
• Nick Fox (Google): “Optimizing for AI search is the same as optimizing for traditional search.”
• Danny Sullivan (Search Liaison): has emphasized that any GEO tools advising content designed “solely for rank and visibility purposes” lose “track of the big picture.”

The position is consistent across the company and over time: SEO fundamentals matter, manipulation will be resisted, there are no shortcuts.

This is the most credible institutional skeptic position because it comes from the platform that benefits most when SEO and GEO converge. Google’s incentive is to keep the optimization disciplines unified under its rules. The position is also empirically partially correct — Bounteous research found 99% of URLs in Google AI Mode appear in the top 20 organic search results. SEO foundations carry through to AI visibility.

Where the platform position over-reaches: it understates the 60% of GEO that is genuinely different from SEO — entity optimization, passage-level structure, cross-platform consistency, earned-media gravity, community presence. These are net-new disciplines that classical SEO does not cover.

The thoughtful counter-position

The honest synthesis: every credible skeptic critique above is partially correct. None of them, taken individually or collectively, justify ignoring the underlying shift.

What the skeptics get right:

• AI engines are not deterministic ranking systems; vendors selling “ranking positions” are misleading buyers.
• Citation drift is real; optimization gains are less durable than SEO equivalents were.
• 80% of GEO overlaps with foundational marketing; agencies pretending otherwise are overstating their differentiation.
• The industry has attracted bad actors; due diligence on vendors is essential.
• SEO fundamentals carry through; teams that ignore them while chasing GEO tactics are building on sand.

What the skeptics get wrong:

• The 20% of GEO that is net-new (entity, passage, cross-platform, earned, community) is where competitive positioning is increasingly decided.
• Citation drift cuts both ways; challengers gain share faster than they ever could against incumbent SEO moats.
• “It’s just good marketing” understates how the structure of marketing has changed when 60% of consumer information starts in an AI conversation.
• The discipline being immature doesn’t mean the underlying shift isn’t real; it means the playbook is still being written.

The honest buyer’s position is somewhere in the middle. Skepticism toward GEO vendors is healthy. Investment in the underlying capability is rational. The question is not whether to engage. It is which vendor, at what scale, with what measurement framework, and on what timeline.

Every statistic in this piece follows Visigent's research standards: named sources, dated figures, corrections on request to help@visigent.com.