The Great GEO Panic Is Over. The Secret, It Turns Out, Was SEO All Along.

There is a certain species of panic that visits digital marketing every few years, dressed in a new acronym and demanding to be taken seriously. This season's arrival goes by GEO — generative engine optimization — and it arrived with the confident swagger of a thing that intends to make your entire career obsolete.

It has, so far, failed to do so. That is more or less the whole story, and it is a delightful one, because the punchline is that the future of search looks suspiciously like its past. Matt Kenyon — a growth partner who spends his days coaxing traffic, leads, and sales out of organic search, and who runs the consultancy Kenyon Digital — has been sent to deliver the news, and he delivers it with the pleasant air of a man defusing a bomb that was never armed. "Ranking in AI search," he says, "pretty much boils down to just doing good SEO — with a few important nuances that can make it easier for tools like ChatGPT to index and present your website."

Read that sentence twice, because it is the entire thesis, and everything that follows is footnotes. The industry, being the industry, has of course produced competing acronyms for the same footnotes. There is GEO. There is AEO — answer engine optimization. There is LLMO, for the completists. "We just love our acronyms here in digital marketing," Kenyon observes, before cheerfully volunteering that his own favorite is KPI, because it "just rolls off the tongue." One admires the honesty. It is rare to watch a discipline gently roast itself mid-explainer.

// The Obituary Was PrematureGoogle Is Not, In Fact, Dead

The strong version of the panic held that generative AI was quietly making Google obsolete. It is a thrilling claim, and it collapses on contact with a spreadsheet. Google, per Exploding Topics, still fields roughly 16.4 billion searches every single day. ChatGPT handles around a billion queries — and, as Kenyon dryly notes, a great many of those are not searches at all but pleas of the "please give me a step-by-step guide on how to like this video on YouTube" variety. You can, he points out, just click the button.

A 2025 study from the agency Higher Visibility adds the human texture: 79.8% of Americans still prefer traditional search engines. People, it emerges, are creatures of habit, and their habit is Google. This does not mean AI search is a mirage — it is unmistakably rising — but the mainstream has not moved, and the smart money is on getting into these tools now, before the crowd arrives. If your customers skew Millennial or Gen Z, Kenyon warns, your runway is shorter: those generations fold AI tools into their search behavior far more readily. The rest of us have a moment. Not a long one, but a moment.

The reassuring corollary is structural. GEO, Kenyon argues, "is less of a new discipline you have to master and more of an additional layer you add onto your existing SEO foundation." When he says "AI search" he means both the generative tools — ChatGPT and its cousins — and Google's own AI Overviews. Different large language models hum under the hood, but the tactics, mercifully, are the same for both.

// Chapter OneWriting the Kind of Content a Machine Falls For

Here is the elegant part. Large language models were raised on conversational human text, and so they gravitate, almost sentimentally, toward content that sounds like a human expert explaining something. The practical instruction writes itself: stop optimizing for the strangled keyword-phrase "best running shoes ranked" and start framing your headings as actual questions a person would ask — "What are the best running shoes for marathon training?" — and then answer them directly, in complete sentences, like a person who knows the answer.

Then comes EEAT, the industry's favorite quadruple — Experience, Expertise, Authoritativeness, and Trustworthiness. A 2024 study found that well-organized, authoritative content, the sort with clean sections and honest FAQ entries, raised inclusion in AI-generated answers by up to 37% on platforms like Perplexity. You demonstrate it, Kenyon says, by sharing first-hand experience and case studies, citing reputable sources, showcasing credentials, presenting multiple perspectives, and — the mark of genuine confidence — addressing your own limitations and objections.

The third instinct is semantic richness. Rather than hammering one keyword until it begs for mercy, weave a web of related terms, concepts, and entities around your topic. This is precisely the guesswork that Surfer's Content Editor is built to abolish — it reads top-performing content and hands you the semantically related keywords, NLP terms, and questions worth covering. And then there is format, where AI reveals itself to be a lazy but honest quoter: it loves a clear hierarchy of headings, it loves bullet points and numbered lists it can lift verbatim, it loves FAQs ("Does your software integrate with Salesforce?" "Yes, our software offers a native Salesforce integration."), and it positively adores a well-built table, from which it will happily pluck a single cell to answer a single question.

// Chapter TwoThe Plumbing: Can the Robot Even Get In?

None of this matters if the crawler is standing at a locked door. So the technical chapter opens, unglamorously, at your robots.txt file — the little text document, findable by typing "robots.txt" after your domain, that quietly tells bots where they may and may not go. Check that you are not accidentally disallowing OpenAI's GPTBot, Google's crawlers, or Microsoft's Bingbot. And here Kenyon delivers the single most enjoyable fact in the entire enterprise: ChatGPT Search runs on Bing's index. Bing. "Humble old Bing," he says, "is making a comeback." So get yourself indexed in Bing and submit a sitemap to Bing Webmaster Tools, a sentence no one expected to be writing in 2026.

The second rule of plumbing: keep your important content in honest HTML text, not buried behind JavaScript or entombed in images and video. Traditional search engines have learned to render JavaScript; AI systems mostly read the raw HTML, and if your key information only materializes after a click, the AI may never see it at all. For the same reason, give your images descriptive alt text and your videos real transcripts. Third: a clean, logical site structure with your best content a few clicks from the homepage, stitched together with descriptive internal links. Which is, Kenyon cannot resist noting, simply SEO best practice wearing a new hat.

// Chapter ThreeSchema, or the Invisible Handshake

If there is a genuinely underused lever here, it is schema markup — a vocabulary of code that humans never see but that machines read to instantly understand what a page is about. It has been an SEO tool for years, and Google now calls it especially important in the age of AI. Organization schema tells the machine about your brand — name, logo, social profiles — planting your entity in the knowledge graphs the AIs consult. Article schema announces publication date, author, and headline, all signals of relevance and credibility. The goal, Kenyon says, is to make every important fact "unambiguous to machines." If an AI can cleanly parse that "ACME Inc is a cybersecurity company founded in 2020, based in NYC," it is far likelier to cite you when someone asks about cybersecurity startups in NYC.

There is no cheat code here, and Google has said so out loud: there is no secret AI tag, no magic trick, to game its results. Schema does not guarantee a rich result or an AI citation — it simply improves the odds, sometimes dramatically, by removing ambiguity. And the facts shouldn't live only on your site; they should be broadcast across the web — Wikidata (the structured spine behind Wikipedia), Google Business Profile, Bing Places, Yelp, Apple Maps, LinkedIn, IMDB, PubMed, Crunchbase, and whatever database rules your particular niche. AI models sample widely, and they trust corroboration.

// Chapter FourAuthority, and the Company You Keep

Finally, trust — which the AI, like a suspicious editor, does not take on faith. It weighs what the rest of the web says about you: quality backlinks from respected sites, mentions in industry publications, positive reviews, and the coveted appearance on "best of" lists. Run a travel blog and land on some well-known outlet's "Top 10 Travel Blogs of 2025," and an AI answering "what are some good travel blogs?" may well repeat your name back to the world. So pitch yourself to the journalists and bloggers who compile those roundups — and, Kenyon insists, bring them something of value: data, an interview, a free trial.

The most durable move, though, is topical authority: covering your niche so thoroughly, from so many angles, that you answer the full spectrum of a reader's questions. The payoff is quietly enormous. Surfer found AI Overviews cite an average of five sources per query, and 90% of the time Google lists eight or fewer — and, crucially, those sources can come from multiple pages on the same domain. Cover a topic completely and you can occupy several of a handful of citation slots at once. Which brings us to the tidy moral of the whole affair, and to the reason a marketer might close this report feeling not doomed but faintly relieved.

And how to know whether any of it is working? Not, Kenyon says, by counting clicks — because in a conversational answer, most people never click a thing. The better metric now is the brand mention: whether the machine says your name. Surfer's newly launched AI Tracker watches for exactly that across popular models, generating relevant prompts and returning a daily report. Which leaves us with the one-line summary the industry spent a whole panic cycle trying to avoid: to be quoted by the machine, be the kind of source a discerning human would quote. The robots, in the end, have excellent taste. It happens to be ours.