Why did Harprit Singh write that AEO and GEO companies are selling repackaged 2016-era SEO?
He was annoyed seeing long-standing SEO concepts rebranded as AEO and GEO tactics, promoted at major conferences without context, which he argues damages businesses that implement them. Rewriting that history means the industry repeats past mistakes.
The post was inspired by a recent San Diego conference where a speaker from an AI search company gave basic SEO advice, such as keyword-in-URL pages and FAQ sections, without caveats. He notes attendees may see short-term gains followed by decreases in both SEO and GEO traffic.
He also questions the value of paying conference fees for this material, pointing to a BrightonSEO recording where a slide claimed pages with keywords in the URL are cited some percentage more times in AI search, a claim he calls trivially obvious.
What do tactics promoted for AEO and GEO actually consist of, according to the discussion?
The tactics being sold are basic SEO staples: building pages that target keywords (relabeled as prompts), putting the keyword in the URL, using keywords in press releases, meta descriptions, FAQs, and paragraphs, even adding a website footer 'for AEO.'
The conversation argues these tactics are important but basic, and that the tactics that work for AEO, GEO, and AI visibility also work for SEO, which remains the foundation: better organic rankings consistently correlate with better AI visibility.
How is the widely cited '85% of pages cited by AI don't belong to you' statistic framed in the discussion?
It is presented as normal search knowledge repackaged to make AI search seem complicated. Page one of Google traditionally has ten links where you own one and not nine, so most cited pages not being yours is unsurprising.
The concern is that these statistics catch the eye of non-SEO executives who then pressure their SEO teams to confirm they are doing specific tactics 'for AI search,' which teams already do. Attribution is also hard because a cited page may rank on page one only for a different keyword than the one being checked.
Who is promoting these AEO statistics and studies, and why is it hard to debate them?
The PR statistics and citation studies generally come from tool companies trying to expand their customer base, while experienced SEO practitioners with deep history are largely aligned against these tactics. Debate is difficult because promoters dismiss critics as 'behind the times' even when the promoted tactics are the old ones.
A guest who joined an AEO accelerator undercover via a burner account found the opening pitch framed AEO as fundamentally different from SEO within the first few minutes, a framing designed to sell tactics packaged as AI search.
Why are AEO and GEO companies expanding beyond AI search optimization?
The AI search optimization category alone is not big enough to deliver investor returns, so companies that raised significant capital are moving into adjacent services such as content engineering, marketing engineering, influencer marketing, and services offerings to grow annual recurring revenue.
Companies named in the discussion include Profound, Aerops, and Searchable, all of which raised substantial funding. The view expressed is that AI search alone will not repay those investors, and the original disruption of SEO was a way to get a foothold in the broader marketing budget.
What does a real enterprise LLM-visibility sprint look like in practice?
A 30-day sprint driven by board demand focused on updating product and solution pages and creating content such as listicles, including one titled 'the best LLMs for finance teams' where the company's product is not placed first, instead positioning honestly around Claude's Excel capabilities and the company's advantages beyond them.
The sprint also included versus pages, which the company had run for paid search for years, and no parasite SEO. For enterprise work, press releases were handled by a third-party PR agency and did not align with the SEO strategy.
How do query fan-outs change what needs optimizing for AI search?
Query fan-outs for commercial B2B terms increasingly include site: searches against Gartner, G2, and Capterra, so work must extend beyond your own website to those third-party profiles. The specific mix differs by company and has to be analyzed per business.
In one financial reporting software example, the fan-out from ChatGPT surfaced Gartner more heavily than expected, prompting a shift to get more reviews and optimize the Gartner profile alongside G2. A practical tactic is ensuring third-party profile descriptions match the brand story on the about page, because AI needs consensus across sources.
The site: colon searches observed came from the initial fan-out, not follow-up research stages, and the LLM was looking for review language on profile pages.
How should brands compete for hyper-competitive keywords like 'best email marketing software'?
You need an optimized solution page, a listicle targeting the 'best' term for Google's AI Overviews, and amplified third-party proof such as G2 grid reports, Trustpilot reviews, and genuine Reddit presence. Dominating requires being in multiple spots on the SERP and across owned, influencer, and press channels.
Recent observations suggest ChatGPT is referencing fewer listicles and more solution pages for comparison terms, though listicles still rank in Google's AI Overviews. Third-party validation matters because the harder question is getting others to call your product the best.
A concrete reputation-fixing example: a client's software implementation was heavily criticized in reviews, so the team published founder social posts and implementation guides stating implementation takes 4 to 8 weeks, and sentiment toward that brand's implementation improved noticeably over roughly two years.
Why is AI search attribution so difficult, and what practical solution is proposed?
Customer journeys involve dozens of steps across prompts and research, making it nearly impossible to attribute AI-influenced revenue to specific tactics or videos. The most basic fix is a 'how did you hear about us' field with an AI option, ideally followed by asking which prompt the customer typed.
Using that button, one company tracked LLM-sourced revenue more than doubling year over year (10 million this quarter with 2 million from LLMs versus 9 million with 200,000 the prior year), which justified the AI search strategy to leadership. Influencer engagement costs of 20,000 to 40,000 dollars are hard to justify to a board when the only metric is citation counts.
Is there a real shift happening from SEO to AEO, or is the language misleading?
Neither speaker believes there is a genuine shift: tactics have always come and gone, and nothing promoted in 2026 as new is actually invented. The phrase 'the shift is shifting' is treated as empty conference rhetoric rather than a real change.
One genuinely newer dynamic discussed is that AI confidently regurgitates claims from press releases that human readers would dismiss, making AI summaries more believable than the underlying promotional material. Claude has started adding disclaimers about self-evaluated vendor lists for terms like best SEO companies, but it cannot yet distinguish a press release on Yahoo Finance from genuine editorial coverage there.
How does small-business AEO differ from enterprise work?
Small businesses should optimize their Google Business Profile and reviews, ensure accurate representation on directories and review sites like OpenTable, Yelp, and free G2 and Capterra profiles, and invest in YouTube videos targeting their keywords, including ungated demos. Small businesses can also get away with aggressive tactics that would attract scrutiny for larger brands.
Brand-reviews websites illustrate the asymmetry: a large brand like Nike buying Nike-reviews.com would draw criticism and links would look suspicious, while a small unknown competitor can build one and AI will pick it up as a third-party source. An AI overview will regurgitate whatever a business makes up if nothing contradicts it.
Why do smaller companies like Gymshark, Orai, Alo Yoga, and New Balance outrank giants like Nike and Adidas on some keywords?
Enterprise SEO at big companies is often very poor because of internal politics, slow processes, and executives who undervalue SEO. New Balance ranks above Nike and Adidas in Canada for men's running shoes largely by executing SEO 101 basics that the giants neglect.
Specific gaps cited: Alo Yoga has pages where the H1 disappears due to JavaScript, Nike's category page copy talks about sustainability and technology with no internal links, and Adidas uses a formulaic internal-link block with heavily branded copy that a brand manager forced the SEO to accept.
New Balance's Canadian page includes the target term in the title and H1, SEO copy with naturally placed internal links, and FAQ-style headings, though the copy itself is generic and AI-written. The lesson for smaller stores is to study what New Balance does and do it better.
What should AEO and GEO companies speaking at big conferences actually be saying?
Three things: acknowledge the history and past effects of the SEO tactics they promote (for example, that Google deprecated FAQ schema), bring actual proof tied to results, and go beyond repackaged SEO with genuinely new ideas such as creator-led AEO.
Even creator-led AEO, where influencers post YouTube videos to appear in AI answers, currently shows citations but no demonstrated link to revenue. Influencer engagements costing tens of thousands of dollars cannot yet be justified to boards without an attribution method.
Will SEO ever die, given AI search?
No. LLMs will never hold all information in training data, so even with their own indexes they must retrieve through relevance and authority signals similar to Google's. Any index an AI company builds will also be easier to game because, unlike Google, it lacks Chrome's behavioral data.
Gaming and patching remain a cat-and-mouse dynamic: fixing one tactic weakens detection of another, so patched tactics eventually work again, and a new model typically takes several cycles to catch a given tactic. As long as people go somewhere to find information, someone will optimize for that place.
Can repositioning a brand up-market, such as from mid-market to enterprise, be done quickly in AI search?
No. A brand with 12 years of mid-market history in the training data faces an enormous effort to become known as an enterprise brand alongside IBM, SAP, Salesforce, and Workday, and the repositioning risks a period where the brand disappears from LLM answers entirely.
Updating old assets with enterprise positioning removes the mid-market signal from the brand's own sources, so it may stop appearing for mid-market queries before earning enterprise visibility. The same dynamic appears after mergers and acquisitions, where the acquired company's previous queries often stop surfacing it.
How can businesses influence AI answers that compare their brand against competitors?
The only way to influence AI is by influencing the sources the AI draws on, which means establishing consensus across as many domains as possible: your own site, third-party review platforms, press releases, YouTube, Reddit, and affiliate listicles.
When a searcher asks AI to compare two brands or choose between two local providers, the AI will produce an answer regardless. Businesses that fail to populate those source channels risk larger companies, third-party websites, and zero-click AI answers eroding their business.