Reading time: 14 minutes
SEO and AI depend on each other, because Google’s AI features use Search’s index and ranking. ChatGPT sometimes queries partner search engines and, according to Microsoft, Bing powers it. So without crawling and indexing, a brand loses ground in both. This guide covers that SEO, not the use of AI to do SEO.
In short
- Google says SEO best practices still apply in AI search. After all, AI Overviews and AI Mode use Search’s index and ranking.
- Clicks got rarer: in an Ahrefs study, the top-ranking page had a 58% lower desktop CTR with an AI summary. However, Seer saw a partial recovery in early 2026.
- Ranking doesn’t guarantee a citation. In an Ahrefs test, for example, only 12% of links cited by AI assistants were in Google’s top 10.
- An allowed crawler, text in the initial HTML and indexing in Google and Bing are the baseline. Finally, Search Console, Bing and GA4 measure the effect.
Does SEO still work with AI Overviews and AI Mode?
It does, and Google says so in writing. AI Overviews and AI Mode retrieve pages from the Search index, based on its ranking systems. What changed is the click: when the answer appears at the top, fewer people visit websites.
Google’s guide to generative AI in Search, updated in July 2026, answers the question right at the start:
“The best practices for SEO continue to be relevant because our generative AI features on Google Search are rooted in our core Search ranking and quality systems.”
Google Search Central (2026)
The guide describes two techniques. In grounding, the ranking systems retrieve up-to-date pages from the index. Additionally, in query fan-out, the model creates related searches to find more sources. How each feature picks its pages is in the guide to Google’s AI Mode and AI Overviews.
How many clicks does SEO lose to AI?
Ahrefs, which sells SEO tools, compared 300,000 keywords using aggregated Search Console data. In December 2025, the presence of an AI Overview correlated with a 58% lower desktop CTR for the top-ranking page. Previously, in the April 2025 study, the drop was 34.5%. In second position, the difference was 50.8%.
On the other hand, Seer Interactive, a search marketing agency, saw a different trend in 2026. Across 53 brands, the organic CTR of searches with an AI Overview rose from 1.3% in December 2025 to 2.4% in February 2026. At the same time, brands cited in the summary got 120% more clicks per impression than those not cited. But Seer warns that this is correlation: strong brands also get more citations.
Without giving numbers, Google’s guide says AI search reaches people who may be more likely to engage and convert. Even so, in practice, the warning sign in Search Console is impressions going up while clicks go down. To separate the summary’s effect from other causes, see what to change when there are many impressions and few clicks. If you need to defend the budget, the 2025 and 2026 numbers are in the analysis of whether SEO is still worth it.
What’s the difference between ranking on Google and being cited by AI?
Ranking means showing up in the results list for the typed search. A citation, in contrast, means the page becomes a source for an answer, often for subquestions the person never typed. A good position helps, but it doesn’t guarantee a citation. After all, AI fans out the question and may use what the web says about the brand.
Ahrefs, for example, put 15,000 long questions to Google, Bing and four AI assistants in 2025. On average, only 12% of the links the assistants cited were in Google’s top 10 for the same question. Perplexity pulled the average up, with 28.6%, while in ChatGPT, Gemini and Copilot the rate was close to 8%. For Ahrefs, the likely explanation is that the question fans out into several searches.
| Criterion | Ranking on Google | Being cited by AI |
|---|---|---|
| What counts | Position in the list | Page used as a source |
| For which search | The typed search | The search and its fan-out queries |
| Where to measure | Performance report | AI reports and fixed prompts |
How does AI change the SEO review?
Work aimed at earning citations got the name GEO, the topic of the guide to what GEO is. In practice, an SEO review for AI gains two questions. Firstly, does the page answer the subquestions AI creates from the search? Secondly, does the brand appear in the third-party sources AI consults?
In fact, what others say about the brand off its site also counts. In an Ahrefs study of 75,000 brands, web mentions of the brand had a 0.664 correlation with presence in AI Overviews. Backlinks, on the other hand, came in at 0.218. It’s correlation, not causation. Even so, it’s worth tracking brand mentions alongside backlinks.
Which parts of technical SEO does AI require?
It requires the basics done well: crawler access, readable text in the HTML and error-free pages. According to Google, the way Search finds and processes pages is the foundation of how its AI accesses content. Outside Google, however, each AI has its own crawlers, and not all of them run JavaScript.
Additionally, the same Google guide says all technical SEO practices remain worthwhile. That includes crawling, a good experience on any device, less duplicate content and care with JavaScript sites.
Crawling: who gets into the site
In ChatGPT, OAI-SearchBot decides whether the site appears in answers that use search. According to OpenAI’s documentation, sites that block this crawler don’t appear in those answers. OpenAI also asks sites to accept its published IP ranges, which involves the CDN and firewall. The guide on how to show up in ChatGPT shows which crawler to allow and which one you can block.
In the audit, therefore, check robots.txt, the CDN and firewall rules and the server logs. For example, the CDN can still block a crawler that robots.txt allows.
Rendering: what the crawler can read
Google renders pages with a version of Chromium and runs JavaScript. Even so, Google’s JavaScript documentation still calls server-side rendering “a great idea” because “not all bots can run JavaScript”.
AI crawlers were in that group in the December 2024 study by Vercel with the consultancy MERJ. None of the major AI crawlers ran JavaScript, including OAI-SearchBot, GPTBot, ClaudeBot and PerplexityBot. Gemini, on the contrary, used Googlebot’s infrastructure and rendered pages. Note that Vercel sells hosting with server-side rendering.
This may be changing, though. On September 30, 2026, a post on DEV Community reported a change in the logs of a Japanese marketplace. According to the author, OAI-SearchBot and GPTBot started rendering pages on September 25, apparently only on some sites. Still, we found no announcement from OpenAI, and its page about its crawlers doesn’t mention JavaScript.
So run a simple test. Open the page’s source code, not the browser’s inspector. If the main text, the price or the title aren’t there, a crawler without JavaScript doesn’t see them. The full checklist, with robots.txt, CDN, logs and rendering, is in the technical SEO audit for the AI era. On React or Next.js sites, see the guides to React SEO and SEO in Next.js.
Why doesn’t your site show up in Google or ChatGPT?
Check the index first, because AI Overviews and AI Mode only cite indexed pages eligible for a snippet. ChatGPT sometimes searches through partners, and Microsoft states that Bing powers it. So a page missing from Google’s index stays out of its AI features. Likewise, if it’s missing from Bing, it loses its chance in ChatGPT.
Google’s guide lists the requirements. Firstly, the page must be in the index and eligible to appear with a snippet. In addition, the site must stay included in generative AI features, a Search Console control that is on by default. But the guide warns: meeting every requirement doesn’t guarantee crawling, indexing or display.
How do you check indexing in Google and Bing?
For Bing, use Bing Webmaster Tools. After all, on the Bing Webmaster blog, Microsoft states that Bing powers AI tools like ChatGPT. The tool is free, imports sites verified in Search Console and processes sitemaps at least once every 24 hours. With IndexNow, the site notifies Bing whenever it creates, changes or removes a page.
A quick diagnosis follows this order:
- Inspect the URL in Search Console. See whether Google has indexed it and whether there’s a noindex, nosnippet or robots.txt block.
- Open the Page indexing report and look for important pages missing from the index.
- Check the Search generative AI control in Settings, in Search Console, and make sure it still includes the site.
- Add the site to Bing Webmaster Tools, submit the sitemap and turn on IndexNow.
- Inspect the same URLs in Bing to confirm they’re in the index.
- Repeat the test prompts after indexing to see whether the page makes it into the answers.
The full diagnosis, error by error, is in the guide on sites that don’t show up in Google or ChatGPT.
Does structured data help you show up in AI?
It helps Google understand the page and the company, but alone it doesn’t guarantee a place in the answer. In fact, Google says generative AI search doesn’t require it and that there’s no special markup for it. It still matters for rich results, products and brand identification.
In the same guide, “Overfocusing on structured data” is on the list of things that aren’t relevant for AI. The passage recommends keeping markup in your SEO strategy, because it helps the page qualify for rich results.
For a brand, structured data is most useful for the entity, that is, telling Google who the company is. According to Google’s organization markup documentation, the markup can help Google understand the company and tell it apart from others. In addition, some properties can influence the logo in results and the knowledge panel.
Google recommends placing the markup on the homepage or the About page. Include the name, logo, contact details and, in the sameAs property, the company’s official profiles. The types that still pay off in Google, and what changes with AI, are in the guide to structured data for SEO.
How do you measure AI’s effect on organic traffic?
With three free tools and your own data collection. Search Console shows impressions in AI Overviews and AI Mode, while Bing Webmaster Tools shows citations in Copilot. Meanwhile, GA4 separates visits from assistants. Finally, to know whether answers cite the brand, the way is to repeat fixed prompts.
On August 31, 2026, Search Console opened the Generative AI performance report to all sites. It shows impressions in AI Overviews and AI Mode, by page, country, device and date. These impressions also go into the regular Performance report, under the Web search type.
According to Microsoft, the AI Performance report in Bing Webmaster Tools entered public preview in February 2026. It shows citations by page in Copilot, in Bing’s AI summaries and in select partners. It also includes a sample of the phrases AI used to retrieve the cited content.
What each source shows about SEO and AI
In GA4, the AI Assistant channel groups visits from sources such as ChatGPT, Gemini and Copilot. However, clicks from AI Overviews and AI Mode stay out of it and go into Organic Search.
| Source | What it shows | What it leaves out |
|---|---|---|
| Search Console (AI report) | Impressions in AI Overviews and AI Mode, by page | Clicks and queries |
| Bing Webmaster Tools | Citations in Copilot and searches AI used | The full list of searches, which comes as a sample |
| GA4 | Visits and conversions from AI assistants | Clicks from AI Mode and AI Overviews |
| Fixed prompts | Brand mentions and citations in each AI | Search volume |
The prompt protocol, with repetitions and margin of error, is in the guide on how to measure brand mentions and citations in AI. The step-by-step for reading AI Mode and AI Overviews data in Search Console is in our guide to the tool.
What does AI change in local SEO and ecommerce SEO?
In both cases, AI reads data that lives outside the site’s text. According to Google’s guide, Google Business Profile and Merchant Center help businesses and products appear in AI answers. In local SEO, that means the profile and reviews, while in ecommerce it means the feed and the product page.
Local SEO
Even with AI, local SEO still starts with your Google Business Profile, with category, hours, photos and reviews. The difference, however, is in the competition. In a SOCi index of chains in the US, AIs recommended far fewer locations than Google’s local 3-pack showed. So name, address and phone number must match across the profile, the site and directories. Each real location also needs its own page.
The numbers and the checklist are in the guide on how to get AI to recommend a local business. What changed in Google’s local pack is in the guide to local SEO in 2026.
Ecommerce SEO
In ecommerce, the catalog reaches Google through the Merchant Center feed and the product markup on the pages. So both must carry the same values for name, price, stock and attributes. In addition, the category page takes on the role of a guide, with price range, use and comparisons between models.
How products get into ChatGPT, AI Mode and Gemini is in the guide to GEO for ecommerce. Categories, product pages and Google Shopping are in the guide to ecommerce SEO.
Where to start reviewing your SEO for AI
Start with what takes the brand out of the running and leave content for later. So the order below goes from crawler access to measurement. It’s for those who do SEO and rank on Google but don’t see their brand in ChatGPT or AI Mode.
- Allow the right crawlers in robots.txt, the CDN and the firewall, and then confirm the visits in the logs.
- Put the main content in the initial HTML, without depending on JavaScript.
- Check indexing in Google and Bing, with a sitemap in both and IndexNow in Bing.
- Review the Search generative AI control in Search Console.
- Mark up the organization with name, logo and official profiles.
- Answer the search’s subquestions on the pages that already rank.
- Earn real mentions off the site, in the press, comparisons and reviews.
- Measure every month in Search Console, Bing, GA4 and with fixed prompts.
What we found when we analyzed 8 pages in Portuguese on SEO and AI
We analyzed the 8 pages in Portuguese that showed up most in 23 searches. Examples include “seo e ia” (“SEO and AI”) and “futuro do seo” (“future of SEO”). Only one of them cites Google’s documentation on generative AI, and even that one is a press release. None mentions Search Console’s AI report, Bing as a source for ChatGPT or OAI-SearchBot.
None covers how AI crawlers read JavaScript, or Google Business Profile and Merchant Center as sources for answers. On the other hand, four recommend structured data to show up in AI, without Google’s caveat that it isn’t necessary. Finally, none gives a study figure on the CTR drop with AI Overviews.
Want to know how your brand shows up in AI answers today?
NotHuman is opening spots for agencies and brands to track how they appear in AI answers. Join the waitlist and we’ll let you know when a spot opens up for you.
Main sources
- Google Search Central, Google’s Guide to Optimizing for Generative AI Features on Google Search (2026).
- Ahrefs, update of the study on how AI Overviews affect CTR (2026).
- Seer Interactive, impact of AI Overviews on CTR, April update (2026).
- Ahrefs, overlap between AI citations and the top 10 in Google and Bing (2025).
- Search Console Help, Generative AI performance report (2026).
- Bing Webmaster, Introducing AI Performance in Bing Webmaster Tools Public Preview (2026).
