Is good AEO just good SEO? What our measurements show

The question comes up in almost every conversation we have with a marketing team: AEO vs SEO (and GEO vs SEO), is answer engine optimization a new discipline, or SEO with a new name and a new invoice? Google says that, for Google, it is "still SEO". Our own measurements of Google AI Overviews, of ChatGPT and of how Microsoft's AI cites our site say the two share a foundation and then split at a step SEO tools never measured. This page sets the two side by side as one stage of the AI search visibility funnel: where good SEO already does the AEO work, where it stops, what a team with a small SEO budget can win, and what to do this week.

Key takeaways

• The foundation is shared. A page that cannot be crawled, indexed and shown with a snippet cannot be used by Google's AI features, and Google's July 2026 guide calls optimizing for them "still SEO".
• The split comes after retrieval. In our study of 150 non-branded US Google searches, about 57% of the websites AI Overviews cited were not in that search's organic top 10 (95% interval 54% to 60%).
• Rank helps, but it does not decide. The first organic result was cited 56% of the time in our data, and results in positions 6 to 10 were still cited 34% of the time.
• A smaller site can win a citation it could never win as a ranking. Answer engines read a handful of sources per question, often from outside page one, so a specific, current answer on a modest site can be read.
• AEO adds work SEO never had: each engine's search crawler, the description an engine gives of you, the third-party pages it reads about you, and measurement that repeats every question.
• Measure it separately. A rank report cannot show citations, and a one-off check of an AI answer is a reading, not a measurement.

Is AEO just SEO with a new name?

No, though the overlap is large enough that the question is fair. Answer engine optimization (AEO) is the work of making a company's pages and facts easy for answer engines such as ChatGPT, Perplexity, Gemini and Google's AI Overviews to find, read, trust and cite, so that the engines describe and recommend the company correctly. SEO is the work of ranking pages in a search engine's results. GEO, generative engine optimization, is a third name for much the same work, from a 2023 research paper; our summary of the GEO paper explains where the term came from.

Google's position, for Google, is that the two are one job. Its page on AI features (last updated December 10, 2025) says: "The best practices for SEO remain relevant for AI features in Google Search (such as AI Overviews and AI Mode). There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." (Google Search Central) Its newer guide (last updated July 10, 2026) goes further: "From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO" (Google Search Central). The newer guide also adds a condition the older page does not mention, covered below.

That is Google speaking about Google. ChatGPT, Perplexity, Claude and Copilot search other indexes, and their vendors publish far less about how they choose sources. And even on Google, what happens to a page after it is retrieved is a different contest from ranking it: a results page shows ten ranked links and lets the buyer choose, while an answer engine reads a few passages and writes one answer that may name you, describe you, compare you with competitors and cite its sources, all before the buyer clicks anything. Success in SEO is a ranking and a visit. Success in AEO is being read, described correctly and recommended, and the visit may never come.

AEO vs GEO vs SEO: what is the difference?

SEO, AEO and GEO share one foundation and differ in what counts as a win. SEO wins a ranked position and a visit; AEO and GEO win a place inside the answer an AI system writes, and the two names describe the same work from two starting points.

In practice most teams use AEO and GEO interchangeably, and so do we on this page. The useful line is the one between both of them and SEO: everything up to retrieval is shared, and everything after it is new. Our summary of the GEO paper shows which rewriting methods the paper tested and how much each moved.

Is GEO different from SEO?

GEO is different from SEO in the same way AEO is: it starts where ranking stops. GEO vs SEO is a question about the unit of success, a sentence inside an AI answer instead of a position on a results page, and the GEO paper measured exactly that.

The GEO paper tested rewrites such as adding quotations and statistics and scored how much of the generated answer each source shaped. Those method names now show up in the searches AI engines run: "When Microsoft's AI cited our pages, 4 of the 73 searches it ran used the GEO paper's own method names, such as 'quotations statistics'." That is our own site, read from Microsoft's grounding-query report: a sign that GEO vocabulary is reaching the searches the engines run.

Is GEO better than SEO? Neither replaces the other. GEO work on a page that cannot be crawled or indexed is wasted, and SEO work that stops at rank leaves the answer to whoever wrote the clearest passage. Treat GEO as the second half of the job SEO starts.

Where does good SEO already do the AEO work?

Good SEO does the AEO work up to the point where the engine has a list of candidate pages. Five things carry over unchanged, and a team that has them in place is already doing half of AEO.

1. Crawlability. The engine's crawler must be allowed to fetch the page. For Google that is still Googlebot, and Google states that "to be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet."
2. Indexing, plus one new setting. A page that is not in the index cannot be retrieved for any question. Google's July 2026 guide adds a site-level condition SEO never had: "a site must be included in Search generative AI features in Search Console to be eligible for display in generative AI features on Google Search." Inclusion is the default, and an owner can switch it off (Search Console Help). Check it once, on the domain property and on any subdomain property.
3. Relevance. The page has to answer the question being asked. In the largest controlled study of what gets cited, 252,000 trials across six language models, topical relevance was one of the two biggest drivers of being cited first (Vishwakarma, Kumar and Jamidar, arXiv 2605.25517).
4. Quality and specifics. The same study found that explicit price information and a recent date helped consistently, while formatting-only edits did little. Pages that state real facts plainly are the pages both kinds of engine can use.
5. Authority built off your site. Links and mentions on credible sites still count. In a study of 112 startups from the 2025 Product Hunt leaderboard, asked about in 2,240 queries, the number of sites linking to a startup predicted whether Perplexity surfaced it, while the startups' scores on a GEO checklist did not (Sharma, arXiv 2601.00912). The author's own summary is to build the SEO foundation first.

Our own Google data shows the same shared base from the other side. Ranking well is associated with being cited:

"The website in the first organic position was cited by the AI Overview 56% of the time (95% interval 51% to 61%). Websites in positions 6 to 10 were cited 34% of the time (31% to 37%)."

That is 393 AI Overview answers on non-branded US searches, counted per website per search. It is an association, not the value of moving up a position: the page in first place is usually also the most relevant and the best known. SaaS searches had the lowest rate at position 1, 42%.

Where do good AEO and good SEO split?

They split after retrieval, when the engine chooses which few passages to read and how to write about you. Our measurements put numbers on five differences.

The answer cites websites the ranking does not show

The clearest number we have is how often the websites an AI Overview cites sit outside the organic results for the same search.

In our words, as cleared for publication: "In our study of 150 non-branded Google searches (US, July 2026), about 57% of the websites cited in AI Overviews did not appear in that same search's top 10 organic results (95% interval 54% to 60%)." 142 of the 150 searches showed an AI Overview. Across a larger panel the gap was nearly universal: "Google showed an AI Overview on 573 of 650 non-branded business searches we tracked. On 91% of the searches we added in July (95% interval 88% to 93%), at least one cited website was missing from the organic top 10."

Independent researchers land in the same place. Kirsten and colleagues (arXiv 2510.11560) found that 53% of the domains AI Overviews used were outside the organic top 10, and 27% were outside the top 100, across 4,706 queries in 2025. Xu, Iqbal and Montgomery (ACM IMC 2026, arXiv 2605.14021) found that nearly 30% of cited domains did not appear anywhere on the first page of results shown with them, across 55,393 trending queries.

The answer is crowded, and the crowd is not all web pages

An AI Overview is not one winner and nine losers. "On 148 commercial SaaS searches such as 'best CRM for small business', Google showed an AI Overview on 143 of them, and each one cited a median of 9 different websites." Across our broader panel, "A Google AI Overview cites about 6 different websites on average, so a single citation is roughly one sixth of the answer."

Some of those sources are not pages a rank tracker follows. "On SaaS searches that showed an AI Overview, YouTube was cited on 69% of them; posts on X on none." That counts searches where YouTube was cited, not what the engine read, and LinkedIn and Facebook pages also appear as sources. For a software company, a clear product video can be a source an SEO plan never budgeted for.

The sources move from day to day

A ranking is fairly stable from one week to the next. An AI answer's source list is not. "When we checked the same 118 searches three times over two days, only 14 of them returned the exact same set of cited websites each time." And "AI Overview source lists turn over fast: across 8 captures in one week, the typical website stayed in an answer's source list for one capture interval." Even whether an AI Overview appears at all moves: "In the same three checks over two days, 22 of 150 searches showed an AI Overview on some checks and not on others (15%; 95% interval 10% to 21%), so a KPI that counts AI Overview appearances can change with nothing changed on your side."

ChatGPT moves too: "Asking ChatGPT the same question three times in a row, the sets of websites it cited overlapped by only about 30% on average, across 100 non-branded questions."

The engine searches for more than the buyer typed

Google says AI Overviews and AI Mode may use "query fan-out", running several related searches across subtopics. Your ranking for the head term is one input among several, and Google does not show site owners those searches. Microsoft does, and our own site shows what they look like: "Microsoft's AI Performance report counted 5,260 citations of our site in Microsoft's AI experiences between March 8 and September 6, 2026, rising from about 3 a day in March to about 63 a day in August. Its grounding-query report listed 73 searches Microsoft's AI ran when it cited us; ten were rewordings of one question." The rewordings were long, specific searches, the kind no one puts in a keyword list. And "When Microsoft's AI cited our pages, 4 of the 73 searches it ran used the GEO paper's own method names, such as 'quotations statistics'." Our site grew over that window, so the rise is not a controlled before and after.

Inside the answer, order and one verifiable fact decide the pick

Once the sources are in front of the model, a ranking signal of a different kind takes over: the order of the list the model reads. In controlled tests where we built the inputs ourselves:

• "In a controlled test with OpenAI's GPT-5.5, when two sources were otherwise equal, the model sided with the one its ranked source list placed first; a single verifiable statement that met the requirement in the question was enough to overturn that preference."
• "In a controlled test with OpenAI's GPT-5.5, adding agreeing sources did not change which company the model picked. What changed the pick was the challenger's position in the ranked list the model was reading."

We have not shown that either holds for other vendors' models. The outside evidence points the same way: in C-SEO Bench (Puerto and colleagues, NeurIPS 2025 Datasets and Benchmarks, arXiv 2506.11097), only 3 of 54 combinations of rewriting tactic and topic produced a significant gain, and a document's position among the retrieved results mattered far more than any rewrite. The practical reading for a page owner: you cannot set the order, but you can make sure the page carries the one specific, checkable statement a buyer's question requires.

The answer describes you, and no ranking measures that

A ranking says nothing about what the buyer reads about you. An AI answer can describe your old product, merge you with a similarly named company, or recommend a competitor while citing your own pricing page. SEO has no metric for that. AEO has to have one.

Can a smaller site win AI citations without a big SEO budget?

Yes, more often than it can win page-one rankings, because the answer draws from a wider pool than the top results. In our data more than half of the cited websites came from outside the organic top 10, results in positions 6 to 10 were still cited about a third of the time, and outside researchers found about a quarter of AI Overview sources outside the top 100. A ranking is a contest for ten slots; an AI answer reads several sources per question, picks passages, and often reaches past page one to find a specific answer.

What wins those citations is not a trick. It is the work this page describes: a page an engine can fetch, an answer stated in the first sentences under a heading that matches the buyer's question, specific facts with dates, and independent sites that say the same thing about you. That work costs far less than a link-building program for a competitive head term, which is why AEO is a real opening for teams that could never outspend incumbents on SEO.

The authority you build still counts. Perplexity surfaced startups with more referring sites more often in the Product Hunt study above, and engines lean on sources they already trust. So the opening is real, and it rewards the same credibility work good SEO always rewarded, pointed at more specific questions.

Does ranking on Google get you cited by ChatGPT?

Somewhat, through the searches ChatGPT runs, but the evidence is observational and the link is far from one to one. The best evidence comes from studies that watched which pages ChatGPT retrieved and which it cited.

• Within ChatGPT, rank and citation travel together. One AEO software vendor ran 15,000 prompts through ChatGPT and recorded 548,534 retrieved pages and 43,233 original and follow-up searches (March 2026). ChatGPT cited only 15% of the pages it retrieved. Pages ranking first in Google were cited 43.2% of the time, 3.5 times the rate for pages ranking beyond Google's top 20, and 32.9% of the cited pages that ranked in a Google top 20 did so only for a follow-up search ChatGPT generated, not for the prompt itself (the vendor's report).
• Across engines, the cited pages barely overlap. A June 2026 preprint asked ChatGPT, Microsoft Copilot, Google and Perplexity the same 15 commercial prompts on one day. The mean overlap between any two engines' cited URLs was 0.0079 on a 0-to-1 scale, and 84.9% of engine pairs shared no cited URL at all (Tannenbaum, arXiv 2609.22655).

Read together: within ChatGPT, pages that rank well in Google for the searches ChatGPT runs are cited more often, so your Google rankings are not wasted there. Across engines, the exact pages cited differ almost completely, so a citation in one engine is no evidence of a citation in another. Neither study shows that moving a page up in Google would cause ChatGPT to cite it, because pages that rank first also tend to be more relevant and better known, and observational data cannot separate those. Treat every "rank X and you get cited Y% of the time" claim, including ours above, as a description of what was seen, not a forecast of what a change will do.

What carries over from SEO, what does not, and what is new?

Most SEO work carries over as the foundation. What changes is the unit of success and how you measure it.

Which "AEO tactics" does Google say you can ignore?

For Google Search, Google's July 2026 guide lists five things you do not need to do for its AI features: create llms.txt files or other "special" markup or Markdown files; break content into small "chunks"; rewrite content just for AI systems; seek inauthentic "mentions"; and overfocus on structured data (Google Search Central). On structured data it says: "Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add." It still recommends structured data for rich results, which our schema markup guide covers.

The popular advice mostly says the opposite. "Of the 8 guide pages we read, 7 recommended at least one of these tactics as a way into AI answers (7 of 8; 95% interval 53% to 98%), usually schema markup, and none of those 7 mentioned Google's statement. The count covers 8 pages chosen because Perplexity cited them for these questions; it is not a survey of the web."

Across engines, the observational evidence on markup is weak too. A 2026 preprint of about 2 million citations from ChatGPT, Claude, Google AI and Gemini, joined to pages from 19 B2B SaaS companies, found that FAQ blocks, structured data and Core Web Vitals looked helpful in pooled data but the effect reversed or fell to zero once each website was compared with itself (Moore and Dunne, arXiv 2609.35077). Google's statements cover Google only, so a tactic Google ignores is untested elsewhere rather than proven useless there. The better use of the budget is the shared foundation above and the pages engines read about you. Our llms.txt guide explains where that file does help: agents reading documentation.

How do you tell whether AEO work is paying off?

Measure AI visibility on its own, per engine and per question, and compare the questions you worked on with questions you left alone. Our guide to what an AI visibility audit actually covers lists what to check first. A rank report cannot show the gap this page describes, and a single check of an AI answer cannot tell a gain from the day-to-day movement above.

The comparison group matters more than it sounds. "In a simulation built on the structure of our own AI Overview data, an analysis that did not know which searches had been worked on never detected an injected improvement, even a 20-point one (0 of 500 simulations at each size; 95% interval 0% to 0.8%). Told which searches had been worked on, the same analysis detected a 5-point improvement in 56% of simulations (95% interval 52% to 60%) and a 20-point one in all 500." In plain terms: write down which questions you are working on before you start, keep a matched set you leave untouched, and judge the work on the difference.

The free reports to start with:

• Google: Search Console's Generative AI performance report shows how often links to your site were shown in AI Overviews and AI Mode (Search Console Help).
• Microsoft: Bing Webmaster Tools' AI Performance report counts citations in Microsoft's AI experiences and lists the grounding searches behind them.
• ChatGPT, Perplexity, Claude and Gemini: no vendor report exists, so ask your buyers' questions repeatedly, on more than one day, and record citations, mentions and whether the description is right. Our guide to prompt research for AEO covers building the question set.

What should an SEO team add for AEO this week?

Start from the SEO you already do and add the steps it never covered. In order, for SaaS and tech companies and for other growing businesses:

1. Keep the technical base clean. Pages indexed, no accidental noindex, answer passages not inside nosnippet or data-nosnippet, sensible canonicals.
2. Open the doors SEO never checked. Allow OAI-SearchBot, Claude-SearchBot and PerplexityBot in robots.txt, confirm your CDN's bot protection lets them through, and confirm the Search Console generative AI setting is on include.
3. Rank for the questions behind your category, not only the head term. Comparison, integration, security, pricing and "best X for Y" pages are where buyer questions and an engine's sub-searches land.
4. Put the answer first. Use the buyer's question as the heading and answer it in the first one or two sentences, with the specifics an engine can lift: price, named integrations, certifications, the date last checked.
5. Keep facts current everywhere you are described. Your site, review listings, documentation and profiles should agree with each other and with what you sell today.
6. Earn independent coverage and publish the formats engines cite. Mentions on credible sites help both channels; a clear product video can be cited where a page is not.
7. Measure AI visibility separately, with a control. Open the Google and Bing reports above, track a fixed question set per engine over repeated asks, and keep a matched set of questions you leave alone.

Frequently asked questions

Should we stop investing in SEO? No. A page has to be crawled, indexed and found to be read at all, and ranked pages are a large part of what AI answers draw on: in our study of Google's AI Overviews, a little under half of the cited websites were in the organic top 10 for the same search, and in ChatGPT, pages ranking first in Google were cited far more often than pages outside the top 20 in one large vendor study. What changes is the target: passages that answer specific buyer questions, measured by whether engines cite and describe you.

Is good AEO the same as good SEO? On Google, Google says optimizing for its AI features is "still SEO", and the foundations are the same. The difference is what happens after retrieval: which passage is read, how you are described and whether you are recommended. Good AEO covers that second half, across every engine your buyers use.

Can a small company get cited by AI without ranking on page one? Yes. In our Google data, about 57% of cited websites were outside the organic top 10 for that search, and positions 6 to 10 were still cited about a third of the time. A specific, current answer to a buyer's question gives a smaller site a real chance.

Do we need a separate AEO team? Usually not at first. The people who own SEO and content can own AEO if they add per-engine measurement of citations and descriptions, and attention to the third-party pages engines read about you.

Does schema markup help with AI answers? For Google, Google says no special markup is needed for its AI features and recommends structured data for rich results. Across engines, the largest observational study we know of found no reliable effect once each website was compared with itself. Keep it for the rich results Google still shows (FAQ rich results are gone; see Google has removed FAQ rich results), and do not expect it to win citations.

Is AEO the same as GEO? In most marketing use, yes. GEO comes from a 2023 research paper and AEO from earlier work on answer boxes and voice assistants; today both describe making a company visible in AI answers.

Is SEO dead? No. In our study of 150 non-branded Google searches, about 57% of the websites AI Overviews cited were outside the organic top 10, which means the rest came from inside it. Pages still have to be crawled, indexed and found before any engine can read them. SEO is the entry ticket; it is no longer the whole game.

What is the equivalent of SEO for AI? Answer engine optimization (AEO), also called generative engine optimization (GEO). It covers the work SEO does not: getting each AI engine's search crawler in, writing passages that answer one buyer question on their own, keeping your facts consistent across the pages engines read about you, and measuring citations per engine.

Is GEO better than SEO? Neither is better; they do different jobs. SEO puts your pages into the pool an AI answer draws from, and GEO decides whether a passage from them ends up in the answer. GEO work on a page no engine can retrieve is wasted, so most teams should run both, with SEO as the base.

How do I optimize content for AEO? Put the buyer's question in the heading and answer it in the first one or two sentences, with specifics an engine can lift: prices, named integrations, certifications and the date last checked. Then allow each engine's search crawler and measure citations per engine. The seven steps above give the order.

Is AEO real? Yes, in the sense that matters to a buyer: AI answers cite a handful of websites per question, often from outside page one, and which ones they cite can be measured and changed. In our Google data, "A Google AI Overview cites about 6 different websites on average, so a single citation is roughly one sixth of the answer," and more than half of the cited websites came from outside the organic top 10.

Work with us

There are two ways to work with us. Teams that want the work done for them can use our managed AEO services: Book a call. Teams that want to see where they stand first can start with Lenz, our free analysis of how ChatGPT and Gemini answer their buyers' questions: Get your free AI visibility report.

How this page was made

Soroush Mehri wrote this page with the help of an AI writing assistant (Claude, made by Anthropic), working from the sources listed below and from our own measurements. Before writing, we read the pages that rank and are cited by Google, AI Overviews and Perplexity for these questions, and set out to cover what they cover and add what they do not. Each factual claim was checked against its source, and findings of ours appear only in wording we have cleared for public use, with the sample size and interval that go with them. Dr. Mehrdad Mehrbod reviewed the page.

Sources

• Google Search Central. "AI features and your website" (last updated December 10, 2025). developers.google.com
• Google Search Central. "Optimizing your website for generative AI features on Google Search" (last updated July 10, 2026). developers.google.com
• Search Console Help. "Search generative AI control" and "Generative AI performance report (Search)." support.google.com, support.google.com
• Kirsten, Grosse Perdekamp, Wu, Upadhyay, Gummadi and Zafar. "Characterizing Web Search in The Age of Generative AI." arXiv 2510.11560
• Xu, Iqbal and Montgomery. "Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact." ACM IMC 2026. arXiv 2605.14021
• Vishwakarma, Kumar and Jamidar. "What Gets Cited: Competitive GEO in AI Answer Engines." 2026. arXiv 2605.25517
• Sharma. "The Discovery Gap: How Product Hunt Startups Vanish in LLM Organic Discovery Queries." 2026. arXiv 2601.00912
• Puerto, Gubri, Green, Oh and Yun. "C-SEO Bench: Does Conversational SEO Work?" NeurIPS 2025 Datasets and Benchmarks. arXiv 2506.11097
• An AEO software vendor. "The Influence of Retrieval, Fan-out, and Google SERPs on ChatGPT Citations." March 2026. Report
• Tannenbaum. "Scoring With the Engine: Retrieval Exposure, Cross-Engine Divergence, and the Limits of Engine-Agnostic GEO Scores." arXiv preprint, 2026. arXiv 2609.22655
• Moore and Dunne. "What Drives Citations in Production Large Language Models? An Observational Multi-Method Study of Two Million AI Citations Across Ten Thousand Web Pages." arXiv preprint, 2026. arXiv 2609.35077
• Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan and Deshpande. "GEO: Generative Engine Optimization." KDD 2024. arXiv 2311.09735
• Elementera, Google AI Overviews panel: 150 non-branded US searches, July 2026 (8 captures, 1,051 answers); 650-search panel; 148 SaaS searches, July 19, 2026; repeat checks of 118 and 150 searches over two days; rank and citation over 393 answers.
• Elementera, ChatGPT repeat-ask measurement over 100 non-branded questions.
• Elementera, controlled tests with OpenAI's GPT-5.5 (inputs built by us), July 2026.
• Elementera, Bing Webmaster Tools AI Performance and grounding-query exports for elementera.com, March 8 to September 6, 2026 (our own site).
• Elementera, review of 8 guide pages Perplexity cited for AEO questions, October 8, 2026, and a simulation over our AI Overview panel's structure (500 runs per condition).

Written bySoroush MehriCo-Founder and Chief AI Officer

Ex-Microsoft. 10+ years of experience in AI/ML, MSc from Université de Montréal (Mila). US patent holder.

Reviewed byDr. Mehrdad "Med" MehrbodCo-Founder and Chief Product Officer

Known professionally as Dr. Med with 12+ years of experience implementing AI products. Strategic advisor to multiple successful founders including a unicorn, US patent holder. Ex-SOTI, UC Berkeley PhD.