What's the difference between AI SEO and traditional SEO?

Traditional SEO targets Googlebot, which can run JavaScript on a delayed second pass. AI SEO targets a different group of crawlers - GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot (Perplexity), and a handful of smaller ones. None of them execute JavaScript. They fetch your URL, parse the initial HTML, and decide whether to cite or train on it from that single response. So a site that's marginal for Googlebot is often invisible to AI search entirely.

Should I focus on traditional SEO or AI SEO first?

Same fix solves both. The pages that don't render content for GPTBot are the same pages Googlebot's web rendering service struggles with. Fixing AI SEO automatically fixes the bot-visibility problem in traditional SEO too, because both ultimately come down to whether your initial HTML response contains your content. Don't think of them as two projects.

Does Google's AI Overview see my site? Isn't Googlebot the same crawler?

AI Overviews use Googlebot for crawling but a different model for synthesis. The crawler still indexes your site the way it always has, but to be cited in an AI Overview, your content needs to be in Google's index AND extractable by the synthesis model. That generally means clean HTML, structured data, and content that lives in the initial response, not in a hydrated component. The same fix that makes you visible to GPTBot makes you eligible for Google AI Overview citations.

What about ChatGPT, Perplexity and Claude - do they have crawlers?

Yes. OpenAI runs GPTBot for ChatGPT search and training. Anthropic runs ClaudeBot. Perplexity runs PerplexityBot. They each fetch URLs to build their indexes and training corpora. None of them execute JavaScript. Each respects robots.txt directives independently, so you can block or allow them per crawler.

What is GEO (Generative Engine Optimisation), and is it the same as AI SEO?

GEO is a newer term coined to describe optimising for generative AI engines specifically. In practice it overlaps heavily with AI SEO - both are about making your content extractable, citable, and groundable for LLM-powered search. GEO tends to put more emphasis on prompt-relevance and entity clarity, but the underlying technical requirements (HTML in the initial response, clean structured data, visible headings) are identical to AI SEO.

How quickly do AI crawlers re-index after I fix this?

Faster than you'd expect. GPTBot re-crawls high-traffic and recently-updated pages weekly or faster. PerplexityBot is even more aggressive - it re-crawls cited pages every few days. ClaudeBot is slower, on the order of weeks. Once you serve rendered HTML to these crawlers, their next visit picks up the change. Most teams using PageGlass see AI visibility improvements within two weeks of switching the CNAME.

Does PageGlass help with AI Overview rankings specifically?

PageGlass makes you visible. Visibility is necessary for citation, but not sufficient on its own. Whether you actually get cited in an AI Overview also depends on topical relevance, freshness, and authority signals - the things any SEO programme should already be working on. We solve the technical floor (your content has to be readable to be considered). The rest of the AI Overview ranking equation is your existing SEO programme's responsibility.

AI SEO: stop being invisible to ChatGPT, Perplexity and AI Overviews

ChatGPT, Claude, and Perplexity can't render JavaScript. Their crawlers fetch your URL, read the initial HTML, and move on. If your site renders content with React, Vue, or any client-side framework, AI search sees nothing - and your site never enters the AI index.

This is a known problem, and the fix is simpler than you think.

Tests your URL the way GPTBot, ClaudeBot, and PerplexityBot do - no JavaScript engine, single-pass HTML fetch.

100+ bots supportedPre-rendered by PageGlassFive-minute setupBuilt on Cloudflare's edge

Don't want an external tool? We understand.

AI search engine optimization isn't a different discipline - it's traditional SEO with a stricter crawler. GPTBot, ClaudeBot, and PerplexityBot are pickier than Googlebot: they fetch once, parse the response, and move on. There is no second-pass JavaScript execution, no patient waiting for hydration. PageGlass renders every URL on the server side for these AI crawlers in five minutes - no SSR migration, no framework lock-in.

Fix this in three steps

Run a free scan. Confirm the bot visibility gap on your URLs. The scan takes ten seconds and needs no signup. You get exactly what GPTBot, ClaudeBot, and PerplexityBot see when they crawl you, side by side with what your visitors see.
Add one CNAME record. Point your domain through PageGlass in your DNS provider. The change takes thirty seconds and is fully reversible. Zero code changes, no deploy, no impact on the bundle your real users download.
Your content is visible to AI search. AI crawlers now receive fully-rendered HTML for every URL on your site. ChatGPT, Claude, and Perplexity can index and cite your content. Come back to PageGlass and run the scan again to confirm the change.

What AI SEO actually is

AI SEO covers the practice of getting your content into the indexes that power AI search products: ChatGPT search, Perplexity, Google AI Overviews, Bing Copilot, Claude with web search, and a growing list of smaller entrants. The vocabulary is still settling - you'll see "AI search engine optimization", "GEO" (Generative Engine Optimisation), "answer-engine optimisation", and a few others used to describe roughly the same thing.

Underneath the terminology, the technical floor is consistent across every AI search product: your content has to be in the initial HTML response when their crawler fetches your URL. Anything that depends on client-side JavaScript - React hydration, async data fetches in useEffect, dynamic component trees - never reaches the AI index.

Why AI crawlers don't run JavaScript

AI crawlers are optimised for crawl volume, not crawl fidelity. GPTBot alone fetches billions of URLs daily across the open web; running a full headless browser on every fetch would multiply infrastructure cost by 10-50x with diminishing returns. So they make a deliberate trade-off: skip the JavaScript engine, accept that some content will be missed, optimise for breadth.

This is different from Googlebot's behaviour. Google has a Web Rendering Service (WRS) that does eventually run JavaScript on most pages, just on a delayed second pass. AI crawlers have no equivalent. There is no second pass; there is the first response, and that's it. If your content isn't there, it isn't in the AI index.

What gets you cited in AI search results

Three things, in order of importance: (1) your content must be in the initial HTML response - this is the technical prerequisite PageGlass solves; (2) the page must be topically relevant to the query - your existing content programme owns this; (3) you should help the model ground on you with clean structured data (JSON-LD, FAQPage, Article schema where appropriate). Some teams also add an llms.txt file at their root summarising their site for LLMs, though support for it is uneven across vendors as of mid-2026.