How to Rank in ChatGPT, Perplexity & Gemini: The 2026 GEO Playbook

Generative engine optimisation (GEO) is the work of getting your brand cited inside the answers that ChatGPT, Perplexity, Gemini and Google’s AI features give your buyers. It isn’t a replacement for SEO; it runs alongside it, and it already brings qualified visitors to the brands that started early. Over the past year, across several B2B and D2C clients, I’ve seen AI referrals grow from almost nothing to a noticeable share of qualified visits. This playbook is what we refined from those experiments as part of our AI SEO services, updated for everything that changed in 2026.

What is generative engine optimisation, and why does it matter in 2026?

Generative engine optimisation is the practice of making your pages easy for AI systems to find, understand, trust and cite. It matters because more and more research now ends inside an AI answer. OpenAI says ChatGPT has more than 1 billion weekly active users, and plenty of them ask it to compare products, services and suppliers.

Google’s AI features work at a similar scale: Google says AI Overviews has more than 2.5 billion monthly active users and AI Mode has passed one billion.

Every one of these engines cites its sources, but clicks are lower. Pew Research found that people clicked a traditional search result in 8% of visits when an AI summary appeared, against 15% when it didn’t. In my client work, the visitors who do click through from AI answers tend to be better qualified and convert well.

Getting cited is not the same as ranking well, and that gap is where GEO comes in. If the term is new to you, start with what GEO is and how it differs from SEO.

What has changed in GEO in 2026?

Quite a lot. Google published official guidance saying that optimising for its AI features is still SEO, free AI reporting arrived in Search Console, Bing Webmaster Tools and GA4, and ChatGPT started showing ads. The core of this playbook holds up, but two tactics I recommended in May, llms.txt and FAQ schema for rich results, no longer earn a place.

• Google’s official guide. Published on 15 May and updated on 10 July 2026, Google’s guide to generative AI features in Search says you can ignore llms.txt, “chunking” and inauthentic mentions, and that structured data isn’t required.
• FAQ rich results are gone. Google stopped showing FAQ rich results on 7 May 2026. FAQ sections still help readers and AI answers.
• Ads in ChatGPT. OpenAI’s August 2026 update says ads run in more than 40 countries, with self-serve buying in India, and that “Advertising does not influence the answers ChatGPT provides.” A citation still has to be earned.

How do generative engines pick their sources?

Each engine runs its own retrieval, but they reward the same basics: content that matches the intent behind a prompt, pages they can crawl and parse, first-hand evidence, a brand they recognise and recent updates. Google adds that its AI features sit on its core ranking systems and run related “fan-out” searches to build each answer.

• Semantic relevance. Match the intent behind the prompt; AI systems understand synonyms and related entities.
• Crawlable, extractable content. Clean HTML and clear headings help engines parse a page, and Google’s AI features only link to indexed pages eligible for a snippet.
• First-hand data and expertise. Original statistics, case studies and author bylines beat recycled SEO content.
• Brand authority. Ahrefs’ study of 75,000 brands found branded web mentions correlated with AI Overview visibility at 0.664, against 0.218 for backlinks.
• Freshness. Genuinely updated content with a clear date helps on time-sensitive queries.

Ranking well no longer guarantees a citation, though. Ahrefs found in March 2026 that only about 38% of pages cited in AI Overviews also ranked in the top 10 for the same query, down from about 76% in mid-2025.

What are the 10 GEO moves that work in 2026?

These are the moves I run for clients now. None of them needs an expensive tool, but they do need you to rewrite content with GEO in mind rather than bolt it on. Compared with the May version of this playbook, llms.txt is out, crawler access is in, and schema is about accuracy rather than rich results.

• 1. Answer the query in the first paragraph. AI answers often quote the opening of a page or section. Google says you needn’t write specially for AI, so treat this as clear writing that is also easy to quote.
• 2. Keep FAQ sections for readers. A short FAQ answers follow-up questions in your buyers’ own words. Just don’t add one to win a rich result.
• 3. Write for entities, not just keywords. Cover related concepts, definitions, examples and edge cases.
• 4. Publish original data. A small 50-sample survey you ran yourself is more citable than a 10,000-sample third-party study you linked to.
• 5. Cite primary sources. Link to official documentation and original research, not to blog posts that summarise them.
• 6. Add author bylines with credentials. A named author with a verifiable bio helps readers and search systems trust the page. Our guide to E-E-A-T after Google’s 2026 core updates shows how.
• 7. Keep structured data accurate. Use Article, Organization and Product markup that matches the visible page. It isn’t required for Google’s AI features, so treat it as tidy labelling, not a ranking trick.
• 8. Let the right crawlers in. Allow OAI-SearchBot and PerplexityBot in robots.txt and make sure Google can index your key pages. Skip llms.txt: SE Ranking’s study of nearly 300,000 domains found no correlation with AI citations. Our comparison of llms.txt and robots.txt explains which file does what.
• 9. Earn genuine brand mentions. Mentions in trusted industry publications, podcasts and communities teach the models your brand is credible. Google warns that seeking inauthentic mentions “isn’t as helpful as it might seem”.
• 10. Update content quarterly. Refresh key pages with new data and a visible updated date.

How does each AI engine differ?

They overlap more than they differ, but each has quirks. In my citation tests, ChatGPT leans on brand recognition, Perplexity on recency and clear answers, Google’s AI features on its core ranking systems, and Claude on authoritative, well-structured pages. Use the notes and table below before you rewrite a page.

• ChatGPT search cites well-known brands far more often than unknown ones, so new brands need mentions first, and the site must allow OAI-SearchBot.
• Perplexity rewards recency and clearly answered questions. Our Perplexity SEO tactics go deeper.
• Google’s AI Overviews and AI Mode draw on Google’s core ranking systems, so the usual SEO signals carry over. See how Google AI Mode picks its cited sources.
• Claude uses a narrower set of sources in my tests and favours pages that read like a reputable publication.

What each AI engine rewards at a glance

How do you measure GEO success?

Track three things: how often you’re cited, how much traffic AI tools send, and whether branded searches rise. This got much easier in 2026. Google, Bing and Google Analytics now report on AI features for free, and a monthly prompt log covers ChatGPT and Perplexity, which still offer no dashboards for site owners.

Search Console’s Generative AI performance report reached all sites on 31 August 2026. It shows impressions from AI Overviews and AI Mode by page, country, device and date, but no clicks.

Bing Webmaster Tools’ AI Performance report shows how often Microsoft Copilot and Bing’s AI summaries cite your pages, and it added a Citation Share view in June 2026.

GA4 now has an AI Assistant channel for visits from ChatGPT, Gemini, Copilot and similar tools. Google’s own AI Overviews and AI Mode are excluded, so those visits still sit in organic search.

For ChatGPT and Perplexity, run 20 to 50 buyer prompts monthly, log which pages get cited and watch branded search volume. Paid trackers help with big prompt sets, but Google’s guide notes that no third-party tool can see inside its ranking or AI systems.

How should Indian B2B and D2C brands start with GEO?

Start with the facts AI tools check first: the same business name, address and offer everywhere, answer-first pages for your main services, and mentions on the Indian directories, marketplaces and industry lists that AI answers already cite. The audience is large: Sam Altman said in February 2026 that India has 100 million weekly active ChatGPT users.

B2B software buyers often ask ChatGPT for shortlists, such as “best CRM for a 50-person sales team in India”. Good SaaS SEO now covers the comparison pages, pricing pages and integration docs that answer those prompts directly.

Consumer brands face product questions instead. In a Google-commissioned Ipsos survey of Indian shoppers who use AI Overviews and AI Mode, 87% said these features help them decide with more confidence. Pages with real specifications, prices and reviews give those answers something to quote, which is why they come first whenever we plan SEO for D2C brands.

If nobody in your team has time to run this, work with an experienced SEO consultant in India who tracks AI citations alongside rankings, not just a monthly keyword report.

How do you run a GEO audit on your top pages?

Start with your ten most important commercial pages and the buyer prompts each one should be cited for. Then check crawler access, rewrite the openings, add evidence and named authors, and log citations every month. The six steps below follow that order, and every one of them also helps your normal Google rankings.

1. List your ten most important commercial pages and the buyer prompts each one should be cited for.
2. Check access: each page is indexed in Google, and robots.txt allows OAI-SearchBot and PerplexityBot.
3. Rewrite the opening so the first paragraph answers the prompt directly, then expand below it.
4. Add first-hand evidence (your own data, photos, prices or examples), a named author with verifiable credentials, and links to primary sources.
5. Keep a short FAQ for real follow-up questions, make sure any structured data matches the page, and work on genuine mentions in trusted industry publications.
6. Run your target prompts monthly in each engine, log citations, check Search Console’s AI report and GA4’s AI Assistant channel, and refresh the pages every quarter.

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, “AI features and your website”

Crawl and indexing problems are the most common blocker I find, and fixing them is standard technical SEO work. For a second pair of eyes, ask for a free SEO audit; it shows the gaps that hold you back in both regular results and AI answers.

The verdict: how do you rank in ChatGPT, Perplexity and Gemini in 2026?

Be easy to crawl, answer the question clearly, prove your experience and get mentioned by sources your buyers trust. Generative engine optimisation adds answer-first writing, citation tracking and genuine mentions on top of good SEO; it doesn’t replace it, and Google now says so in writing.

In most niches I audit, few competitors are doing this properly yet. Audit your ten most important pages against this playbook, measure every month and keep going.

Frequently Asked Questions