At Search Central Live Tokyo on 29 May 2024, Google Search Relations analyst Gary Illyes confirmed that AI Overviews and the experimental AI Mode use the same crawling, indexing, and ranking systems that power traditional Google Search results. In short, the practices that help pages rank in standard search also surface them in these generative features—no separate “AI search optimization” is required.
How AI Overviews Ranks Content
Illyes said Googlebot, the core index, and long-standing ranking signals determine what appears in AI Overviews. Google assesses page quality without checking whether the content was produced by a human or an AI system.
The generative pipeline then:
- Employs RankBrain and the Multitask Unified Model (MUM) to interpret intent in 75 languages.
- Runs a “query fan-out” step to suggest follow-up searches.
- Uses “grounding” checks to verify answers against indexed pages.
Event Details
- Date: 29 May 2024
- Event: Search Central Live Tokyo
- Speaker: Gary Illyes, Analyst, Google Search Relations
- Key quote: “AI Search Optimization = SEO.”
Background
RankBrain, introduced in 2015, helps Google handle unfamiliar queries. MUM, added in 2021, connects information across text, images, and video. Both models now operate throughout crawling, indexing, and ranking. AI Overviews entered public testing through Search Labs in 2023 and rolled out to U.S. results in May 2024.
Sources
Further details are available in Kenichi Suzuki’s LinkedIn summary and on the Google Search Central Live event page.
.svg)
.svg)
.svg)