Four-index reality
The four-index reality is the structural fact that AI search isn't a single discipline because five major AI search engines source from four distinct underlying indexes — Google (for AIO + Gemini), Bing (for ChatGPT), Brave (for Claude), and Perplexity's own crawler.
Definition
The four-index reality is the framework Citare uses to describe why AI search visibility cannot be optimized as a single uniform target. The five major generative AI search engines that customers actually use ground their responses on four distinct web indexes:
- Google index — feeds AI Overview, AI Mode, and Gemini (confirmed by Google's 2026-05-15 AI Optimization Guide)
- Bing index — feeds ChatGPT web search
- Brave index — feeds Claude when grounding via Brave's API
- Perplexity's own crawler index — feeds Perplexity
Why it matters
Different indexes mean different crawl priorities, different ranking signals, and different content quality assessments. A brand can surface heavily in one and be effectively invisible in another for the same query. Industry data shows surface rates for the same brand on the same query regularly differ by 30-40 percentage points across platforms.
What it implies for measurement
- Single-source AI search tools that only measure ChatGPT (or only AIO) report on one of four indexes. They miss the visibility picture across the other three.
- Measurement has to be platform-by-platform, not pooled. Aggregate "AI surface rate" numbers obscure the gap that matters.
- Tactics that work on one index can produce zero lift on another. llms.txt, for example, is respected by Claude/Perplexity/OpenAI but explicitly declined by Google as of 2026-05-15.
What it implies for tactics
- For Google surfaces: classic SEO foundations. Helpful content, structured data, page experience, technical hygiene.
- For Bing-backed ChatGPT: Bing Webmaster Tools indexing + IndexNow auto-pings + content that aligns with how Bing ranks (more weight on entity disambiguation than Google).
- For Claude on Brave: Brave's index is far smaller and recency-biased. Fresh content earns disproportionate citation.
- For Perplexity: PerplexityBot is a dual-horizon crawler — same crawl event feeds training data and live retrieval. Server-side rendering matters more here than anywhere else.
See /four-index-reality for the full archetypes + diagnostic patterns and /four-platform-reality for the user-experience companion view.
Frequently asked
Why four indexes and not five if there are five major AI search platforms?
Google AI Overview and Google Gemini share Google's index — they're two products on the same crawl + ranking infrastructure. ChatGPT, Claude, and Perplexity each ground on different indexes (Bing, Brave, and their own crawler respectively). So five platforms, four distinct underlying indexes.
Did Google's 2026-05-15 guide invalidate the four-index reality?
No. Google's guide confirms one of the four indexes — the Google index used by AIO and Gemini. The other three indexes (Bing for ChatGPT, Brave for Claude, Perplexity's own) are not addressed by Google's guide and remain structurally separate. The four-index reality is more correct after 2026-05-15, not less.
How much can surface rates differ across the four indexes for the same brand?
30-40 percentage points is common for the same brand on the same query in the same week. The variance is structural, not measurement noise — different crawlers, different ranking signals, different audiences using each engine.
Related
Stop guessing where you rank in AI search
Citare measures citation rate and share of voice across ChatGPT, Google AI Overview, Gemini, Claude, and Perplexity — weekly, for your priority queries. Free forever tier.