How to Optimize Your Website for AI Search

That last sentence is the part people do not want to hear, so let me be straight up front. The tactics look a lot like SEO. Three things genuinely do not: how you measure it, how you research prompts, and the fact that Google is only one of the places you need to show up. Everything else is work you probably already know how to do. If the definitions are still fuzzy, here is where GEO and SEO actually diverge.

Here is the order I would do it in.

Step 1: Let the crawlers in

Nothing else matters if the bots cannot read your pages.

Open your robots.txt and look for the AI crawlers by name. OpenAI runs four. OAI-SearchBot surfaces your site in ChatGPT's search answers. GPTBot collects content for training. ChatGPT-User handles requests a person triggers, and OAI-AdsBot checks ad landing pages. Blocking GPTBot keeps you out of training data but leaves ChatGPT search alone. Blocking OAI-SearchBot takes you out of ChatGPT search results. A lot of sites blocked both years ago and never went back to look.

Google's documentation is blunt about the rest: there are no additional requirements to appear in AI Overviews or AI Mode. Your page needs to be indexed, eligible to show with a snippet, and crawlable. The one thing Google does stress is that important content has to exist as text. If your page builds its main content with JavaScript and the raw HTML is nearly empty, that is a real gate.

One more box. Since 31 August 2026, every site worldwide has a Search generative AI control in Search Console that keeps your content out of Google's AI features. Confirm nobody on your team switched it on. Opting out means no impressions and no traffic from those features, though Google says it is not a ranking factor for ordinary search.

Step 2: Read what Google actually published

In May 2026 Google published a guide on optimizing for generative AI features. It is worth twenty minutes, mostly for what it tells you to stop doing.

Google's position is that "optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Then it lists what you do not need: new machine readable files, AI text files, markup, or Markdown versions of your pages. You do not need to break content into tiny pieces for AI, or write in a special way for generative engines. Structured data is not required. Chasing inauthentic mentions does not help.

Not everyone agrees. Ethan Lazuk, writing the day after the guide went up, kept the llms.txt file on his own site, arguing there is no real downside since Claude and ChatGPT can read one. Fair enough. My read is that llms.txt costs an hour and has no published evidence behind it, so treat it as a cheap experiment, not a step.

Step 3: Improve the page, not the tricks

This is where most AI search advice falls apart.

The original 2024 GEO paper found that adding quotations, statistics and citations lifted visibility by double digits, and most checklists online are built on it. A newer paper pushes back. Liu and Xu, in April 2026, tested those same tactics on GPT-4o-mini and found every one scored below an untouched page. The unmodified baseline hit 13.34% visibility. Statistics dropped it to 12.21%, quotations to 11.08%, citing sources to 11.78%, an authoritative tone to 11.94%. Their own method, which reworks the whole document rather than bolting on phrases, reached 18.31%.

These two papers disagree and nobody has reconciled them. The second one's explanation rings true: the additions "introduce redundancy and disrupt the naturalness signals that generative engines implicitly favor." Stuffing a statistic into a thin page gives you a thin page with a statistic in it.

So write the page properly. Say something that is not already on ten other pages. Put the answer near the top. Use headings that match real questions. Keep sentences short enough to lift cleanly into an answer. Link your pages to each other.

On schema, the evidence is reasonably clear. Ahrefs, a vendor with products in this space, tracked 1,885 pages that added JSON-LD between August 2025 and March 2026 against 4,000 control pages. Citations barely moved: down 4.6% in AI Overviews, up 2.4% in AI Mode, up 2.2% in ChatGPT, the last two inside the noise. Their caveat matters. They studied pages already cited heavily, so this says nothing about pages starting from zero. Otterly, also a vendor, tested seven platforms and found six could not fetch or correctly read schema when asked. Add schema because it helps you in ordinary search, keep it matching the visible text, and do not expect it to move AI citations.

Step 4: Turn measurement on, then learn to distrust it

Search Console now has a Generative AI performance report covering AI Overviews and AI Mode, rolled out worldwide on 31 August 2026. It is the only first-party number you will get.

Then understand what it withholds. It reports impressions. No clicks, no click-through rate, no query data. An impression means Google showed a link to your site inside an AI feature. It does not mean anybody visited you.

August gave everyone a lesson in reading it carefully. Starting 13 August 2026, a logging error made impressions in that report fall off a cliff. Practitioners across LinkedIn spent a week comparing charts and assuming their AI visibility had collapsed. John Mueller confirmed it was "just a logging issue and not representative of visibility changes in Search." Glenn Gabe later noted the bug had briefly made it possible to calculate a real click-through rate for AI features, since the ordinary Performance report carries that data too. Nobody ran the numbers in time.

The practical rule: before you change anything based on a drop in that report, check your server logs for AI crawler hits and compare against ordinary organic clicks for the same URLs. If crawls held steady, the dashboard moved and you did not.

Step 5: Look outside Google

Here is the gap the Search Console report cannot close. It is Google only. It tells you nothing about ChatGPT, Perplexity, Claude or Copilot.

So build a small prompt set, twenty or thirty questions a real customer would ask, and run them by hand across the platforms you care about. Record who gets named and which pages get quoted. Do it monthly, on the same day. It is slow and unglamorous and it is the only way to see platforms that report nothing.

Watch your server logs too. Crawler hits from GPTBot, PerplexityBot and Google-Extended tell you who is reading you, and they move before the dashboards do.

One caveat worth knowing. In August 2026, Lily Ray noticed large volumes of site: queries in Search Console with thousands of impressions and almost no clicks, and suggested they might be ChatGPT fan-out queries landing in Google's data. She published a regex for checking your own sites. It is unconfirmed, and her own commenters pointed at rank trackers as an equally good explanation. Treat it as a curiosity, not a metric.

Where to start tomorrow

Check robots.txt. Confirm you have not opted out in Search Console. Pick your five most important pages and make each genuinely better than what it competes against. Watch the Generative AI report for a month before concluding anything. Then build your prompt set for the platforms Google cannot show you. If you would rather not squeeze that between other work, that is what we do.

Sources

• Google Search Central, Optimizing your website for generative AI features on Google Search, published May 2026, developers.google.com/search/docs/fundamentals/ai-optimization-guide
• Google Search Central, AI features and your website, developers.google.com/search/docs/appearance/ai-features
• Google Search Central Blog, Introducing Search Generative AI performance reports in Search Console, June 2026
• Search Engine Land, Google Search Console AI performance reports and Search generative AI control rolling out globally, 31 August 2026
• Search Console Help, Generative AI performance report (Search), support.google.com/webmasters/answer/16984139
• OpenAI, Bots and crawlers documentation, developers.openai.com/api/docs/bots
• Zikang Liu and Peilan Xu, Think Before Writing: Feature-Level Multi-Objective Optimization for Generative Citation Visibility, arXiv 2604.19113, 21 April 2026
• Louise Linehan and Xibeijia Guan, We Tracked 1,885 Pages Adding Schema. AI Citations Barely Moved, Ahrefs, 11 May 2026 (vendor research)
• Rick Tousseyn, Schema Markup's Real Impact on AI Search, Otterly.ai, 23 March 2026 (vendor research)
• Ethan Lazuk, What I Think About Google's Optimizing Your Website for Generative AI Features Guidance, 19 May 2026
• Raja Abu Bakar, LinkedIn, Google Search Console Generative AI Performance Report data bug, 17 August 2026, quoting John Mueller
• Glenn Gabe, LinkedIn, on the missed CTR calculation during the Search Console reporting bug, 22 August 2026
• Lily Ray, LinkedIn, on suspected ChatGPT fan-out queries appearing in Search Console, 12 and 31 August 2026
• Search Engine Journal, Google's New AI Search Guide Calls AEO And GEO 'Still SEO'