Field Guide · Gemini
How Gemini Chooses Sources. The Google citation model: AI Overviews on the SERP, Gemini chat, and the search infrastructure underneath.
Gemini sits on top of the same search infrastructure that drives Google ranking — so the citation signals that work in classic SEO mostly work for Gemini, with added weighting for entity clarity, schema depth, and AI Overview eligibility. The signals are familiar; the surfaces are new.
See live citation rates across all four engines at the Bttr. Citation Index →
Citation signals
The signals that move Gemini.
Each engine weights signals differently. The list below ranks the ones that actually move citation rate for Gemini, with the relative weight we observe in measurement.
High weight
Google search ranking
Pages that rank top-3 for the underlying query are dramatically more likely to be cited in AI Overviews and Gemini chat. Classic SEO is the foundation.
High weight
Schema markup depth
Article, FAQPage, HowTo, Person, Organization schema. Gemini reads structured data directly and the AI Overview block heavily reuses entities extracted from schema.
High weight
Entity clarity (Knowledge Graph)
Pages that map cleanly to Google Knowledge Graph entities (with sameAs links, Organization/Person nodes, clear topical category) earn AI Overview citations more reliably.
High weight
AI Overview eligibility
Not every query triggers an AI Overview. Queries that do — informational, definitional, comparison — are the citation surface. Pages targeting Overview-eligible queries are the leveraged targets.
Medium weight
Freshness
For time-sensitive queries, Gemini surfaces recent content fast. For evergreen queries, depth and authority outweigh recency.
Medium weight
Core Web Vitals
Performance metrics still matter as a ranking signal, which feeds AI Overview eligibility. Underrated leverage.
Medium weight
Backlink graph
Backlinks contribute to underlying Google ranking, which feeds Gemini. Less direct than schema but still load-bearing.
Medium weight
Google-Extended opt-in
If a site has opted out of Google-Extended (the Gemini training crawler), training-cycle citations are blocked. AI Overview citations may still fire from live retrieval. Most sites benefit from opting in.
Retrieval mode
Two surfaces, same engine underneath.
Gemini powers two distinct citation surfaces: AI Overviews on the Google SERP (the block above the blue links for eligible queries) and Gemini chat (the standalone chat product). Both pull from the same retrieval layer — Google search index plus the Knowledge Graph — but the routing logic differs.
AI Overviews fire for informational queries where Google judges synthesis adds value. Citation slots are limited (typically 3-7 sources per Overview), and the cited URLs are roughly the top-10 organic results for the underlying query, re-ranked for entity clarity and schema depth.
Gemini chat is more like ChatGPT or Claude: longer-form synthesis, broader context, citation block included. The same content that earns AI Overview citation tends to earn Gemini chat citation.
The implication for AI Search Visibility: classic SEO still matters, but schema + entity clarity are now the marginal levers. Sites with strong SEO that have not invested in schema depth are leaving Gemini citation on the table.
What moves citation rate for Gemini.
- Top-3 Google ranking + schema depth. The leveraged combination. Top-3 organic ranking earns AI Overview eligibility; deep schema markup wins the citation slot inside it.
- Knowledge Graph entity alignment. Organization, Person, and Place schema with sameAs links lifting your entity into the Knowledge Graph. AI Overviews disproportionately cite Knowledge-Graph-recognized entities.
- AI Overview-targeted pillar content. Definitional, comparison, "how to" pillar pages are the formats that trigger AI Overviews. Sales pages do not.
- Strong Core Web Vitals. Underrated. Performance feeds underlying Google ranking, which feeds AI Overview eligibility. A page that ranks 7th won't earn an Overview citation; the same page ranking 3rd will.
What does not work.
- Schema without ranking. Schema alone does not earn AI Overview citation if the underlying page is not ranking top-10. Schema amplifies ranking; it does not substitute for it.
- Google-Extended opt-out. Opting out of Google-Extended blocks training-cycle Gemini citations. Most brands benefit from opting in unless there's a specific reason not to.
- Sales pages targeting informational queries. AI Overviews fire on informational intent. Sales pages get downweighted for those queries, even when keyword overlap is strong.
Frequently asked
Gemini, common questions.
Is Gemini just an AI layer on Google search?
Effectively yes — Gemini AI Overviews use the same retrieval layer as classic Google search, with re-ranking for entity clarity and schema depth. Strong SEO is the foundation; schema and entity clarity are the marginal levers.
How do I know if my query triggers an AI Overview?
Run the query on Google. If an AI Overview block appears above the blue links, the query is Overview-eligible. Most informational, definitional, and "how to" queries are. Most transactional queries are not.
Does Google-Extended opt-in matter?
Yes for Gemini training-cycle citations. Opting out blocks Anthropic-style training citation. AI Overviews can still cite via live retrieval, but the long-term compounding is gated. Most brands benefit from opting in.
How fast does Gemini reflect new content?
For AI Overviews: roughly as fast as Google ranking moves — days to weeks for fresh authoritative content. For Gemini chat: similar timing, slightly more conservative on freshness.
What is the most leveraged Gemini optimization?
Top-3 organic ranking + deep schema markup on the page. That combination wins AI Overview citation slots reliably. Most other moves are secondary.
The other engines
Read the matching deep dives.
Bttr. Field Brief
More on Gemini and the other three engines.
Monthly. One signal worth your time on Brand Operating Systems, AI search visibility, and the infrastructure buildout. No filler.