The one-picture mental model

EngineRetrieval baseFreshnessCitation styleFirst optimization move
Google AI Overviews / AI ModeGoogle index + Knowledge GraphIndex cadenceLink cards beside the answerClassic SEO health + answer blocks
ChatGPTBing index + OAI-SearchBot live fetch + training memoryMixedSources panelBing indexation + allow OAI-SearchBot
GeminiGoogle index via grounding with Google SearchIndex cadenceInline links, Workspace surfacesSame as AIO + Google-Extended access
PerplexityLive retrieval (PerplexityBot)Real-timeInline numbered sources, alwaysFast, structured, fresh pages
CopilotBing index + enterprise graph (M365)Index cadenceInline linksBing Webmaster + IndexNow
ClaudeClaudeBot crawl + opt-in web searchMixedSparse, primary-source preferencePrimary documents, clean authorship
GrokWeb + real-time X firehoseReal-time (X)Posts + linksX-native presence

Read the deep-dives

Foundation

How LLMs choose sources

The shared logic: chunk-level retrieval, corroboration, authority priors, and why consistency beats volume.

Read →
2.5B+ monthly users

Google AI Overviews & AI Mode

Query fan-out, Gemini 3 grounding, trigger logic, and what the 2026 AI Mode merger changed.

Read →
~900M weekly users

ChatGPT

Three crawlers, three jobs. Bing's gatekeeping role, the sources panel, and training-data memory effects.

Read →
750M+ app MAU

Gemini

Grounding with Google Search, Google-Extended controls, and the Workspace surfaces brands forget.

Read →
1B+ queries/mo

Perplexity

The retrieval-first purist: live fetches, inline receipts, freshness bias, and the fastest citation market.

Read →
Enterprise surface

Microsoft Copilot

Bing-grounded answers inside Windows and M365 — where B2B research quietly happens.

Read →
Researcher favorite

Claude

Anthropic's careful sourcer: sparse citations, primary-document preference, outsized analyst influence.

Read →
Real-time contrarian

Grok

xAI's engine reads the live X conversation — the only major surface where social velocity is a ranking input.

Read →

How to use this section

Read the foundation chapter first, then the engine(s) where your buyers actually are — the free audit tells you which those are. Each engine chapter follows the same structure: pipeline mechanics, crawler policies, citation behavior, optimization checklist, measurement notes.

Frequently asked questions

Which AI engine should I optimize for first?
The one your buyers use, measured not guessed. As defaults: consumer categories start with Google AI Overviews and Gemini (largest reach); B2B starts with ChatGPT and Perplexity (procurement research behavior); enterprise software adds Copilot. The free audit shows where you currently stand on each.
Do engines share source preferences?
Partially. All engines reward crawlability, direct answers, entity consistency and corroborated facts. They diverge on infrastructure (whose index), freshness weighting, and how much training memory substitutes for retrieval — which is why per-engine playbooks exist.
How fast do engine behaviors change?
Constantly at the margins, with step changes at model releases (e.g. Gemini 3 becoming the AI Overviews default in January 2026). We update these chapters on material changes — check the updated date on each page.