Google's AI Overviews -- the AI-generated summaries at the top of Google search results -- started as an experiment called Search Generative Experience (SGE) in May 2023. After two years of testing, Google retired the SGE name and launched AI Overviews as a standard feature for all users in May 2024. If you read any early guides about optimizing for SGE, the core advice still applies -- the content signals did not change dramatically in the transition, though the model and some interface details did.
The most important thing to know for beginners is that AI Overviews reaches far more users than SGE did. SGE required users to opt in through Google's Search Labs program; AI Overviews is on by default for all users in supported regions. This means the AEO opportunity is significantly larger. Content that earned SGE citations generally continued earning AI Overview citations -- there was no major content strategy reset required for most publishers.
The three practical beginner takeaways: any content you created to earn SGE citations is still valid for AI Overviews. Update your article titles, headings, and internal documentation to use the term 'AI Overviews' instead of 'SGE' now that the product name has changed. And add named author credentials to your key content pages -- AI Overviews' stronger EEAT enforcement (vs SGE) makes author attribution more important than it was during the experimental period.
SGE to AI Overviews: A Complete Timeline of Changes
Google's transition from Search Generative Experience (SGE) to AI Overviews changed the citation signals, query coverage, and ranking models in multiple stages between May 2023 and Q1 2026. Each transition shifted which content properties mattered most. Understanding the history explains why current AEO best practices exist.
SGE vs AI Overviews: What Actually Changed for Content Publishers
Most SEO content written during the SGE era (2023 to May 2024) remains compatible with AI Overviews -- the core content signals did not change. But six specific aspects of the system changed in ways that require content strategy updates. Hover any row.
| Feature | SGE | AI Overviews |
|---|---|---|
| Availability | Opt-in (Search Labs enrollment required) | Default ON for all users in supported regions |
| Model | PaLM 2 (synthesis) + Google's search retrieval | Gemini Pro / Flash (synthesis) + Google's search retrieval |
| Citation chips | Horizontal carousel of 3 top sources | Expandable source list (3 visible, more on expand) |
| YMYL trigger | Very restricted -- most YMYL queries excluded | Slightly more YMYL coverage with stronger disclaimers |
| Schema signals | FAQPage dominant; Article secondary | FAQPage + Article + Person (author) all significant |
| Freshness | Moderate freshness weighting | Higher freshness weighting (dateModified emphasized) |
Adapting SGE-Era Content for AI Overviews: Four Concrete Examples
Content optimized for SGE in 2023 needs targeted upgrades for AI Overviews in 2026. The changes affect author attribution, passage structure, FAQ format, and data sourcing. Toggle between before and after to see the exact improvements needed.
Published by: Editorial Team | Tags: AI, SEO
No named author, no credentials. Works for traditional SEO but fails AI Overviews EEAT filter.
Background: Understanding AI search is complex. There are many factors. Historically, search engines have evolved from keyword matching to semantic understanding to the current era of generative AI. All of these changes affect how content is discovered.
Context-first, buried answer. Poor cross-encoder re-ranking score. No direct answer in first sentence.
Frequently Asked Questions: General questions about our service...
Non-specific FAQ title with no actual questions. FAQPage schema added to this structure provides zero citation signal benefit.
Recent studies show that AI search is growing rapidly and that most marketers are adjusting their content strategies in response.
Vague attribution, no named source. Lower cross-encoder quality score for both SGE and AI Overviews. Not citable.