Indexing: definition and why AI answers depend on it

99Visibility Research · Updated October 21, 2026 · 4 min read

Short answer

Indexing is a search engine storing a page it has fetched so it can retrieve that page later for a query. Each engine keeps its own search index. Google, Bing and the indexes behind ChatGPT search and Perplexity are separate, so a page indexed in one can be missing from another. AI answers that use live retrieval can only draw on pages held in the index they search.

Key takeaways

Indexing means a fetched page is stored and can be retrieved later, which is different from being crawled or being cited.
Every engine has its own index, so healthy index status in Google tells you nothing about Bing, and the reverse.
Google says a page must be indexed and eligible for a snippet to appear as a supporting link in AI Overviews or AI Mode.
Being indexed is necessary for retrieval-based AI answers but does not guarantee a citation.
Check index status engine by engine, and treat noindex, snippet controls and JavaScript-only content as the usual reasons a page never gets stored or shown.

What indexing means

A crawler fetches a page. Indexing is the next step: the engine processes what it fetched and stores it in its search index, a database it can query when someone asks a question. A page that was crawled but not indexed cannot be returned for any query.

The key point is that there is no single index. Google keeps one. Bing keeps one. Perplexity runs its own crawler, PerplexityBot, for search indexing. OpenAI uses OAI-SearchBot to surface sites in ChatGPT search. Your index status is a set of answers, one per engine, not one fact.

How indexing works

Three things have to go right for a page to be stored.

The crawler can reach it. robots.txt must allow that engine's bot, and your firewall or CDN must not return a 403 to it.
The crawler can read it. Vercel found that ChatGPT and Claude crawlers fetch JavaScript files but don't execute them, so text that appears only after JavaScript runs may never be stored. Googlebot does render JavaScript.
The page is eligible. A noindex directive tells the engine not to keep the page:

<meta name="robots" content="noindex">

Part of indexing is never visible from outside. The engine decides what to keep, and Google states plainly that "indexing and serving isn't guaranteed." For the common technical blockers, see what code-level issues stop AI search engines from indexing a page.

Why indexing matters for AI answers

When an AI engine searches the web before answering, it pulls candidates from an index. If your page is not in that index, it cannot be a candidate.

Google's documentation, as of October 2026, says a page must be "indexed and eligible to be shown in Google Search with a snippet" to appear as a supporting link in AI Overviews or AI Mode, with "no additional technical requirements."

Otterly's 2026 experiment with two sites of AI-generated posts showed the per-engine split clearly. Google deindexed both sites, and Google's AI surfaces almost stopped citing them. Bing kept both sites indexed, and Microsoft Copilot, which grounds on Bing, kept citing them: 2,718 of the 3,934 citations Otterly tracked for one site. To see which AI answers draw on which index, read Bing index vs Google index.

The next move is simple: check each engine separately using the steps in checking index status engine by engine.

Common misconceptions

"Indexed means cited." No. Indexing makes a page eligible to be retrieved. Which retrieved pages get used is the model's decision, and nobody outside the engine controls it.
"Indexed in Bing means visible in ChatGPT." In Otterly's test, the sites ranked in Bing and drew effectively zero ChatGPT citations, even with ChatGPT's crawlers visiting.
"robots.txt allow means indexed." Allowing a bot buys eligibility only. A blocking firewall, a noindex tag or JavaScript-only content can still keep the page out.
"Pinging IndexNow gets you indexed." The IndexNow documentation guarantees receipt of the ping, not indexing, and Google does not consume it.
"Blocking GPTBot removes you from ChatGPT search." OpenAI says each bot setting is independent. GPTBot governs training. OAI-SearchBot governs search.

Frequently asked questions

How do I check my index status?

Use each engine's own tools. Google Search Console shows whether Google has indexed a URL, and Bing Webmaster Tools does the same for Bing. Neither tells you about the other. There is no public tool that shows whether ChatGPT or Perplexity has stored a page. The [free index status check](/free/free-index-status-check) on this site covers Google and Bing.

Can a page be indexed but still not appear in AI Overviews?

Yes. Google requires a page to be indexed and eligible for a snippet. A nosnippet or max-snippet:0 directive keeps an indexed page out of the answer text. Even with full eligibility, Google says serving isn't guaranteed, and AI Overviews often do not trigger for a query at all. Eligibility is in your control. Selection is not.

Why do AI answers still cite a page I deleted?

An engine can hold a stored copy until it recrawls the URL and drops it, and different engines refresh on different schedules. Answers given without live retrieval rely on what the model learned before its knowledge cutoff, which no index update changes. See [why AI answers still cite pages you deleted](/questions/ai-cites-deleted-pages) for the fixes that apply.

Does structured data help a page get indexed?

Google says structured data isn't required for generative AI search, and an Ahrefs matched-control study found adding schema produced no major uplift in AI citations. Existing markup does no harm, but it is not a lever for AI answers. Indexing depends on access, readable HTML and eligibility directives.

Sources

This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.