AI-generated answers now sit at the top of many Google searches, and more people ask ChatGPT, Gemini or Perplexity instead of searching at all. That has changed how much traffic a good ranking brings in, and it has created a market of new acronyms and “AI optimisation” tools. This guide separates what has actually changed from what hasn’t, drawing on Google’s own guidance and independent research, and sets out what’s worth doing in 2026.
What has actually changed
Fewer clicks from the same rankings
When Google shows an AI Overview, people click through to websites less often. A Pew Research Center study of real browsing behaviour found that users clicked a traditional search result in 8% of visits where an AI summary appeared, compared with 15% where it didn’t. Only about 1% of visits included a click on a link inside the summary itself. Industry studies since then have found similar or larger drops for top-ranking pages.
The effect isn’t the same for every search. Short factual questions (“what is…”, “how many…”) lose the most clicks, because the answer is on the results page. Searches with commercial or local intent (“accountant near me”, “buy…”, “best… for…”) still send people to websites, because people need to compare options, check prices and get in touch.
New places to be found
AI assistants cite sources. When ChatGPT search, Perplexity or Google’s AI Mode recommends a product, a service or an answer, it usually links to the pages it drew from. Being one of those cited sources is a new kind of visibility, and it’s measurable, because visits from chatgpt.com or perplexity.ai show up in your analytics as referral traffic.
New controls and reports from Google
- A Generative AI performance report in Search Console (introduced in June 2026) shows how often your pages appear in AI Overviews, AI Mode and similar features, broken down by page, country and device. It reports impressions, not clicks.
- An opt-out setting in Search Console (rolled out worldwide by the end of August 2026) lets site owners choose whether their content appears in, and helps ground, Google’s generative AI features. Google says using it doesn’t affect ordinary rankings. It also means no traffic from those features.
What hasn’t changed
Google’s own guidance on its AI features is blunt: they are “rooted in our core Search ranking and quality systems”, so the best practices for SEO remain relevant. Industry studies have consistently found that the pages cited in AI Overviews are mostly pages that already rank well in ordinary results.
Google also lists things you don’t need to do:
- No
llms.txtfile. Google says Search doesn’t use them, and they neither help nor harm. - No “chunking”. You don’t need to break content into tiny pieces for AI to understand it.
- No special writing style just for AI search.
- No special structured data. Structured data is useful in general, but not required for AI features.
Treat tools and agencies that sell “AI search optimisation” as a separate discipline with some caution, especially any that claim special access to Google’s data.
What’s worth doing in 2026
1. Get the technical basics right
AI features draw on the same index as ordinary search. If a page can’t be crawled, isn’t indexed, or hides its main content behind scripts, it can’t appear anywhere. Check Search Console’s indexing report, fix errors, keep pages fast on mobile, and make sure important content is in the HTML rather than loaded later.
2. Publish what can’t be found elsewhere
Google’s guidance puts it directly: don’t “recycle what others on the internet have already said, or could easily be produced by a generative AI model”. Generic content is exactly what AI summaries replace. What they can’t replace is first-hand material: your own pricing and process, real examples, original data, expert explanations of your specific field, and clear answers to questions your customers actually ask you.
3. Make trust visible
- Say who is behind the content: your company, its details and how to contact it.
- Cite primary sources for factual claims, and date your content so readers and systems can see it’s current.
- Keep company information consistent across your website, Google Business Profile and directories.
4. Answer questions clearly, then go deeper
Put a direct answer near the top of a page, and then give the detail, the exceptions and the “how” below it. This isn’t a special AI format. It’s simply good writing, and it helps both readers and the systems that summarise pages.
5. Double down on commercial and local searches
Service pages, pricing pages, comparison pages and location pages match searches where people still click. For local businesses, a complete and active Google Business Profile matters more than ever, because Google’s AI features draw on it too.
6. Be present beyond your own website
AI answers draw on reviews, forums, video platforms, industry publications and directories as well as company websites. Genuine reviews, useful answers in communities where your customers ask questions, and coverage in relevant publications all contribute. Avoid anything inauthentic, such as fake reviews or paid mentions dressed up as opinions.
Decide which AI systems can use your content
Several crawlers now visit websites for different purposes, and blocking the wrong one can remove you from places you want to appear:
| Crawler or control | What it affects |
|---|---|
Googlebot | Google Search, including AI Overviews and AI Mode. Blocking it removes you from Google entirely. |
| Search Console AI features setting | Whether your content appears in Google’s generative AI features, without affecting ordinary results |
Google-Extended (robots.txt) | Use of your content for Gemini model training and some Gemini grounding. It does not remove you from AI Overviews. |
OAI-SearchBot | Whether your pages can appear in ChatGPT search answers |
GPTBot | Whether your content may be used to train OpenAI’s models |
For most businesses that want to be found, the sensible default is to allow search crawlers (including OAI-SearchBot). Then decide separately, as a business question, whether to allow model-training crawlers.
Measure differently
- Report enquiries and sales from organic search, not just traffic. Traffic to informational pages may fall while enquiries hold steady.
- Watch impressions alongside clicks in Search Console. Rising impressions with falling clicks usually means AI features are answering the question on the results page.
- Use the Generative AI performance report to see which pages appear in AI features.
- Track AI referrals by adding a segment in your analytics for visits from
chatgpt.com,perplexity.ai,gemini.google.comand similar sources. - Ask new clients how they found you. “ChatGPT recommended you” is an answer you can only capture by asking.
A short checklist
- Important pages are indexed, fast on mobile, and have their main content in the HTML.
- Service, pricing and location pages answer commercial questions in specific terms.
- Content includes first-hand detail that a generic summary can’t provide.
- Company details are consistent across the website, Google Business Profile and directories.
- robots.txt allows the search crawlers you want, and the training crawlers you’ve decided to allow.
- Reports track organic enquiries, AI feature impressions and AI referral traffic.
Sources and further reading
- Google Search Central: Guide to optimizing for generative AI features
- Google Search Central: AI features and your website
- Google: New opportunities, control and insights for website owners
- Search Console Help: Generative AI performance report
- Pew Research Center: Do people click on links in Google AI summaries?
- OpenAI: Overview of OpenAI crawlers
Search features and crawler policies change often. This article reflects the position as of September 2026.