GEO (generative engine optimisation) is the work of getting your brand named, cited and recommended inside AI answers: Google’s AI Overviews and AI Mode, ChatGPT, Perplexity and the rest. SEO is the work of ranking pages in search results and earning the click. They are not rivals. GEO sits on top of SEO, and most of it is SEO done properly.
I run both for clients at Straight Up Search, mostly eCommerce and SaaS brands, and the question I get most is some version of “should we stop paying for SEO and do GEO instead?” The short answer is no. The longer answer is below, including the parts that genuinely are different and where I’d put the money.
Does GEO replace SEO?
No. Google has said so in writing. Its own guide to AI features states: “From Google Search’s perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.” (Google Search Central)
The mechanics back that up. Google says there are “no additional requirements to appear in AI Overviews or AI Mode”, and that to be shown as a supporting link a page must be indexed and eligible to show in normal Search with a snippet (Google Search Central, AI features). If a page can’t rank, it can’t be cited. Outside Google the same logic applies: ChatGPT’s search features rely on OpenAI’s own crawler, and OpenAI says sites that block OAI-SearchBot “will not be shown in ChatGPT search answers” (OpenAI). Crawlability, indexation and authority are SEO jobs, and every AI engine depends on them.
What GEO changes is the finish line. SEO asks “do we rank?” GEO asks “when someone asks an AI assistant about our category, does it mention us, link to us, and say something accurate?” Those are related questions with different answers, which is why GEO is worth treating as its own workstream rather than a rebrand.
Where GEO and SEO overlap
More than the hype suggests. These all serve both:
- Technical foundations. Pages that can be crawled, rendered and indexed. An AI engine can’t quote a page its crawler never saw.
- Authority. Links and brand mentions from sites people trust still shape which sources get picked.
- Helpful, specific content. Pages that answer the question directly and back it up.
- Clear site structure. Sensible internal linking so the important pages are easy to find.
If your SEO is in poor shape, fix that first. GEO work on a site with indexing problems is money spent in the wrong order.
Where they genuinely differ
| SEO | GEO | |
|---|---|---|
| Goal | Rank a page and earn the click | Get named, cited or recommended in the answer |
| Unit of success | Position, clicks, organic revenue | Mentions, citations, how you’re described |
| What you optimise for | Keywords and search intent | Prompts and the sub-questions behind them |
| Where you show up | Your own page in a list of results | A sentence in someone else’s summary, often without a link |
| What counts | Mostly your own site | Your site plus what reviews, forums, comparison pages and press say about you |
| Measurement | Mature: Search Console, rank trackers, analytics | Immature: prompt sampling, citation tracking, lots of noise |
Three of those differences matter most in practice.
Prompts fan out. Google says AI Overviews and AI Mode “may use a ‘query fan-out’ technique”, running several related searches across subtopics to build one answer (Google). So a single question can pull from pages ranking for a dozen different searches. Covering the follow-up questions around a topic matters more than squeezing one page towards position one.
Off-site counts more. When someone asks “what’s the best X for Y”, the answer is usually assembled from reviews, Reddit threads, comparison articles and listicles, not just brand sites. What third parties say about you becomes part of your visibility. I covered one version of this, the AI mentioning your brand but linking to someone else, in AI Overviews and eCommerce search.
The click is less certain. Pew Research tracked 900 US adults’ real Google searches in March 2025. When an AI summary appeared, people clicked a traditional result in 8% of visits, against 15% when no summary appeared, and clicked a link inside the summary in just 1% of visits (Pew Research Center). Bain & Company’s consumer research found about 80% of search users rely on AI summaries at least 40% of the time (Bain & Company). Being mentioned in the answer is now part of the job, even when it doesn’t send a visit.
What the research says works for GEO
The term GEO comes from a 2023 paper by researchers at Princeton and IIT Delhi, later published at KDD 2024 (Aggarwal et al., “GEO: Generative Engine Optimization”). They rewrote source pages in different ways and measured how visible each version became in generated answers. Their headline finding was that the right changes could boost visibility by up to 40%.
The useful part is which changes worked:
- Adding quotations, statistics and cited sources performed best, with relative improvements of roughly 28 to 41% on their main visibility measure.
- Making text clearer and more readable also helped, by around 15 to 30%.
- Keyword stuffing gave little to no improvement, and on Perplexity it did worse than doing nothing.
Two caveats before anyone rewrites their whole site off the back of it. It was a controlled experiment on a benchmark set of queries, not a study of live Google rankings, and the authors themselves found results varied by subject area. Treat it as direction, not a formula. The direction is sensible anyway: specific, well-sourced pages with real numbers are easier for an AI system to quote than vague marketing copy.
GEO myths worth ignoring
A lot of GEO advice sold right now contradicts what Google has published. From the same Google guide:
- “You need an llms.txt file.” Google says you don’t need new machine-readable files, AI text files, markup or Markdown to appear in its AI features, because “Google Search itself doesn’t use them.” Other AI companies may treat these files differently, but for Google it’s not a lever.
- “Chop your content into small chunks for AI.” Google: “There’s no requirement to break your content into tiny pieces for AI to better understand it.”
- “There’s special AI schema.” Google: “Structured data isn’t required for generative AI search, and there’s no special schema.org markup you need to add.” Normal structured data is still worth having for rich results, it just isn’t an AI switch.
If an agency’s GEO pitch is mostly those three things, that tells you something.
How to measure GEO (honestly)
This is the weakest part of GEO today, and it’s where I’d push back hardest on anyone selling guaranteed outcomes.
- Track a fixed set of prompts. Pick the 15 to 30 questions your buyers actually ask, run them across Google AI Mode, ChatGPT and Perplexity on a schedule, and record whether you’re named, linked, or recommended.
- Separate mentions from citations. Being named without a link and being linked as a source are different outcomes. Track both.
- Check what’s said, not just whether. Wrong prices, old product names or a competitor’s feature attributed to you are all fixable, but only if you look.
- Expect noise. AI answers change from run to run. Look at trends over weeks, not single screenshots.
- Keep SEO reporting separate. Google counts AI Overview and AI Mode traffic inside the normal Web search totals in Search Console (Google), so don’t expect a tidy “GEO traffic” line in your clicks report.
For a sense of how differently the engines behave, the ChatGPT vs Google AI research we wrote up is a good starting point. The same brand can do well in one and be missing from the other.
How I’d split the effort
There’s no universal ratio, and anyone giving you one without looking at your site is guessing. This is how I think about it:
- Weak technical SEO or thin content: almost everything goes to SEO. GEO depends on it.
- Solid SEO, research-heavy buying journey (SaaS, high-ticket eCommerce, anything people compare before buying): add GEO as a real workstream. Comparison content, third-party reviews, accurate product and pricing information, and being present where your category gets discussed.
- Solid SEO, impulse or low-consideration products: SEO and product feeds still do most of the work. Monitor AI answers for accuracy and keep GEO light.
In every case, the same pages usually serve both. A good comparison page ranks in Google and gets quoted by ChatGPT. That’s why I’d rather run them as one plan with two sets of measurements than as two separate budgets fighting each other.
If you want help with the AI side, our generative engine optimisation service covers it, and it’s built on the same foundations as our SEO service.
FAQ
Is GEO replacing SEO?
No. Google’s own guidance says optimising for its generative AI features is “still SEO”, and AI engines rely on pages being crawlable and indexed. GEO adds new goals (being mentioned and cited in answers) and new measurement, but it sits on top of SEO rather than replacing it.
Is SEO dead now with AI?
No. Clicks from some searches are falling where AI summaries appear, and Pew Research found people click results less often when a summary is shown. But AI answers are built from web pages that have to be found, indexed and trusted first. What’s changing is how you measure success, not whether search visibility matters.
Is SEO still worth it in 2026?
Yes, for most businesses that sell online. Organic search still drives traffic and revenue, and the same work (technical health, helpful content, authority) is what makes you eligible to be cited in AI answers. The case gets stronger, not weaker, if you care about AI visibility.
What’s the difference between GEO and AEO?
Mostly the name. AEO (answer engine optimisation) was used for featured snippets and voice assistants, and is now often used for AI answers too. GEO usually refers specifically to generative AI tools like ChatGPT, Perplexity and Google’s AI Overviews. In practice the work overlaps almost entirely.
Can I do GEO without SEO?
Not really. You can build mentions on third-party sites, but your own pages still need to be crawlable, indexable and trusted to be cited. Doing GEO on a site with SEO problems means paying for the second step before the first.
