AI Search Optimization for Every Platform
AI search optimization is not one job. ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity and Claude get their answers in different ways, cite in different ways and are reached by different crawlers, so a page can be visible on one and missing on the next.
The levers that hold up under scrutiny are about access, and they are subtractive: let the right crawler in, do not hide your text from it, keep your facts readable without JavaScript. OpenAI's and Anthropic's crawlers fetch pages without running JavaScript (Vercel, “The rise of the AI crawler”, December 2024); Googlebot renders it.
What does not hold up is the additive kind. The best-designed test of rewriting content for AI citation found most methods ineffective and often harmful (C-SEO Bench, arXiv:2506.11097), and Google's AI optimization guide says there is no special markup to add.
Below, each platform as 99Visibility measures it: how we ask it, whether it searches the web, its weight in the score, and the crawlers that reach it, with what the vendor says happens when you block them. It comes from the same configuration the audit runs on.
ChatGPT
- How we ask it
- OpenAI API, gpt-6-luna, with web search
- Web search
- On: it can read the web when it answers, so a page that is reachable and readable can change what it says.
- Weight in the score
- 36% of platform coverage
- Crawler that fetches
- OAI-SearchBot (ChatGPT search). Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links. Source: developers.openai.com
- Training crawler
- GPTBot. Your content is excluded from training OpenAI foundation models. This does not affect ChatGPT search — OpenAI states each setting is independent. Source: developers.openai.com
Google AI Overviews
- How we ask it
- Read from Google's results page in your market
- Web search
- On: it can read the web when it answers, so a page that is reachable and readable can change what it says.
- Weight in the score
- 18% of platform coverage
- Crawler that fetches
- Googlebot (Google Search, AI Overviews, AI Mode). Pages drop out of Google Search. AI Overviews and AI Mode are built into Search and use the same crawler — Google states there is no separate robots.txt token for them. Source: developers.google.com
Google AI Mode
- How we ask it
- Read from Google's AI Mode in your market
- Web search
- On: it can read the web when it answers, so a page that is reachable and readable can change what it says.
- Weight in the score
- 10% of platform coverage
- Crawler that fetches
- Googlebot (Google Search, AI Overviews, AI Mode). Pages drop out of Google Search. AI Overviews and AI Mode are built into Search and use the same crawler — Google states there is no separate robots.txt token for them. Source: developers.google.com
Gemini
- How we ask it
- Gemini API, gemini-3.8-flash, grounded on Google Search
- Web search
- On: it can read the web when it answers, so a page that is reachable and readable can change what it says.
- Weight in the score
- 9% of platform coverage
- Crawler that fetches
- Googlebot (Google Search, AI Overviews, AI Mode). Pages drop out of Google Search. AI Overviews and AI Mode are built into Search and use the same crawler — Google states there is no separate robots.txt token for them. Source: developers.google.com
- Training crawler
- Google-Extended. Your content is excluded from training future Gemini models and from Vertex AI grounding. Google states it does not impact inclusion in Google Search nor is it used as a ranking signal. Source: developers.google.com
Perplexity
- How we ask it
- Perplexity API, perplexity/sonar, with web search
- Web search
- On: it can read the web when it answers, so a page that is reachable and readable can change what it says.
- Weight in the score
- 18% of platform coverage
- Crawler that fetches
- PerplexityBot (Perplexity search index). Your pages leave the Perplexity search index. Perplexity states this crawler is not used to crawl content for AI foundation models — it is search only. Source: docs.perplexity.ai
Claude
- How we ask it
- Anthropic API, claude-haiku-5-5, with web search
- Web search
- On: it can read the web when it answers, so a page that is reachable and readable can change what it says.
- Weight in the score
- 9% of platform coverage
- Crawler that fetches
- Claude-SearchBot (Claude search). Anthropic says this agent improves search result quality; blocking it likely keeps your pages out of the results Claude search draws from. Anthropic does not state the consequence outright. Inferred from: support.claude.com
- Training crawler
- ClaudeBot. Your content is excluded from Anthropic training crawls. Anthropic runs search (Claude-SearchBot) and user fetches (Claude-User) as separate agents. Source: support.claude.com
What every platform has in common
Allowing a crawler buys eligibility, not a mention. It means the engine may fetch your page; whether it then uses it is the engine's call.
Blocking a training crawler is not the same as blocking search. OpenAI states that each of its settings is independent, so you can keep GPTBot out and still let OAI-SearchBot in.
robots.txt is half the gate. A firewall or bot-management rule can refuse a crawler that robots.txt allows, and nothing in robots.txt will tell you. The AI crawler access check in every 99Visibility plan, Free included, reads both.
See where you stand on each platform
The Free plan audits ChatGPT and Google AI Overviews; paid plans add the rest. Read how multi-platform coverage works, or run the free AI visibility checker.
Start for Free