By Yanisa Chubplun, Integrated Communications Manager, Foundeast
Maybe GEO vs SEO is the wrong fight.
Read enough marketing content this year, and GEO sounds like a changing of the guard: SEO out, GEO in. But much of what makes a page discoverable inside ChatGPT or Google’s AI Overviews still rests on work SEO teams have done for years — technical structure, useful writing, a site that can be crawled and understood. The more useful question is what changes when search stops pointing people towards pages and tries to answer the question itself.
Search is still search. The interface has changed.
Google says this plainly in its own guidance: SEO best practices remain relevant because its generative AI features are rooted in the same core ranking and quality systems as standard search. AI Overviews and AI Mode still rely on Google’s Search index and core ranking systems, so pages need to be crawlable, indexed, and eligible to appear in Search. A page still has to be crawlable, indexed, and eligible to appear with a snippet.
GEO doesn’t give a brand permission to forget SEO. It changes what can happen after SEO has done its job. If a page isn’t indexed or useful, no amount of “AI-friendly” formatting can guarantee inclusion in a generative search response.
One search is no longer necessarily one query
Google says AI Overviews and AI Mode may use query fan-out: one question can trigger several related searches before the system assembles a response. An example is a search for how to fix a lawn full of weeds — the system might also pull in herbicide options, chemical-free methods, and prevention, folding it into one response.
That reframes the working question for content teams. Traditional SEO thinking asks what keyword we’re ranking for. Answer-search thinking asks something broader: what larger question is the system resolving, and which part can our content help with?
Read as “build a page for every sub-query,” and it’s the wrong lesson. Google warns against exactly that, calling it ineffective and a risk under its scaled-content policies. Depth beats a scatter of thin pages chasing fan-out fragments.
The unit of visibility has changed
Classic SEO measurement runs on a familiar sequence: page, ranking, impression, click. Answer-based search introduces a parallel one — information contributes to a synthesised answer, which may or may not carry a citation, which may or may not lead anywhere else.
“Ranking a page and contributing to an answer are related goals. They are no longer the same goal.”
A 2026 University of Toronto study, accepted at the EDBT/ICDT 2026 Workshops, found meaningful differences between the source domains used by traditional web search and generative AI systems, including differences in source type and freshness. It’s one study rather than a universal rule, but it reinforces the point: appearing in an AI-generated answer and ranking prominently in traditional search aren’t necessarily the same outcome.
Different answer engines, different rules
It’s a mistake to treat “AI search” as one system with one rulebook. Google is explicit that no special AI markup is required — no llms.txt file, no new schema, no chopping content into machine-readable fragments.
OpenAI takes a different approach: OAI-SearchBot surfaces websites in ChatGPT’s search results, while GPTBot separately crawls content that may be used in model training. Publishers can control the two independently through robots.txt — allowing search visibility while opting out of training.
There’s no single universal AI-search infrastructure — one discoverability strategy, applied differently depending on which system a brand needs to reach.
Measurement needs a wider lens
Rankings, impressions, clicks, organic traffic and conversions still matter, and Search Console reports on them — now including a Generative AI performance report for visibility inside AI Overviews and AI Mode. ChatGPT, meanwhile, adds utm_source=chatgpt.com to referral links, giving publishers a way to identify traffic that clicks through. But referral data only shows visits that happen — not every appearance or influence inside an AI-generated answer.
The question is no longer only where we rank. It’s also where we contribute to the answer, and what happens next.
One search strategy, not two competing ones
None of this argues for separate SEO and GEO operations, or two teams chasing the same ground. It argues for one discoverability strategy stretched across a wider set of surfaces. SEO still supplies most of the foundation. A GEO lens adds a few extra questions worth asking of any content plan: can the systems that matter access it, does it hold up when one question becomes several, and where — beyond a ranking — might the brand show up?
The organisations that navigate this well won’t be the ones that drop SEO fastest. They’ll be the ones that know which foundations still hold, which assumptions no longer do, and where visibility has widened.
Foundeast works with brands across Thailand on search and content strategy that spans both — from technical groundwork through to how a business shows up as AI search grows into how people look for things. If your current approach doesn’t yet account for both, that’s worth a conversation.
Source List
- Google Search Central — Optimizing your website for generative AI features on Google Search. Last updated 10 July 2026. https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
- Google Search Central — AI features and your website. https://developers.google.com/search/docs/appearance/ai-features
- Google Search Central Help — Generative AI performance report (Search Console). https://support.google.com/webmasters/answer/16984139
- OpenAI Developers — Overview of OpenAI Crawlers. https://developers.openai.com/api/docs/bots
- OpenAI Help Center — Publishers and Developers – FAQ. Page shows “Updated 14 days ago” as of retrieval on 14 August 2026 (approx. late July 2026). https://help.openai.com/en/articles/12627856-publishers-and-developers-faq
- Chen, M., Wang, X., Chen, K., and Koudas, N. (2026). Navigating the Shift: A Comparative Analysis of Web Search and Generative AI Response Generation. Accepted at EDBT/ICDT 2026 Workshops (Tampere, Finland). Preprint submitted 23 January 2026, revised 16 May 2026. University of Toronto. https://arxiv.org/abs/2601.16858