Search Generative Experience (SGE) was Google's experimental generative AI search feature, tested in Search Labs from May 2023. It placed an AI-generated snapshot above the traditional results, synthesizing information from multiple web pages, linking to sources and allowing conversational follow-up questions. In May 2024, Google replaced the experiment with AI Overviews, later joined by AI Mode. SGE is now a historical name for Google's first generative AI search interface.
Practical example
Someone searching "best CRM for a small business" with SGE enabled saw an AI-written snapshot above the blue links. It listed a few platforms, explained what each suited and showed link cards to the pages it drew from. A software vendor's comparison guide could be cited there even when it was not the top organic result, or be skipped despite ranking well. AI Overviews works the same way today.
Why is SGE important for AEO?
SGE was the first large-scale demonstration of an answer engine inside Google Search. It established the pattern that answer engine optimization (AEO) and generative engine optimization (GEO) now address: an AI synthesizes an answer and chooses which pages to cite, so visibility depends on being selected as a source, not only on ranking position.
Users can get their answer in the snapshot without clicking a result, which changed how traffic and brand exposure are measured. Guidance written for SGE largely still applies to AI Overviews and AI Mode, which Google says are rooted in its core ranking and quality systems and need no special markup or optimization.
How did SGE work?
SGE added a generative layer on top of Google's existing search systems rather than replacing them. The process had four stages:
- Triggering. Google's systems decided whether a query warranted a generative response. The feature was deliberately limited to certain query types, with guardrails for sensitive topics.
- Retrieval. The same ranking and quality systems that power classic Search identified relevant web pages for the query.
- Generation. A large language model wrote an "AI-powered snapshot" summarizing the key points from those pages.
- Presentation. The snapshot appeared above the organic results with links to the sources it drew on, an expandable view showing how the response was corroborated, and suggested follow-up questions. Tapping a follow-up opened a conversational mode that carried context from one question to the next. Ads continued to appear in dedicated, labeled slots.
This retrieval-then-generation approach is the same basic mechanism that AI Overviews and AI Mode use today, with newer Gemini models and techniques such as query fan-out.
From SGE to AI Overviews and AI Mode
| Date | Milestone |
|---|---|
| 10 May 2023 | Google announces SGE at Google I/O as an opt-in experiment in Search Labs, available in the US in English on Chrome desktop and the Google app. |
| Late 2023 | The experiment expands to additional countries and languages, still opt-in through Search Labs. |
| 14 May 2024 | Google launches AI Overviews to all US users and says people had already used the feature billions of times through the Labs experiment. SGE as a named product ends here. |
| 5 March 2025 | Google introduces AI Mode as a Labs experiment: a conversational search mode for complex, multi-part questions, built on a custom Gemini model and a query fan-out technique. It opened to all US users in May 2025. |
| Early 2026 | Gemini 3 becomes the default model for AI Overviews, and users can continue from an AI Overview into an AI Mode conversation. |
| October 2026 | Google updates both surfaces with more prominent inline links, a section highlighting official websites and expanded news carousels. |
In practice, "SGE" in older articles means what "AI Overviews" means today. When discussing Google's current AI search features, use the current product names.
Related concepts
- AI Overviews: the production successor to SGE. The core experience is the same: an AI-generated summary with links, shown only when Google's systems judge it more helpful than classic results. It has a far wider reach than SGE ever had, since no opt-in is required.
- AI Mode: a separate, conversational search mode for exploration, comparison and multi-step questions. Google states that AI Mode and AI Overviews may use different models and techniques, so the links they cite can differ.
- Query fan-out: the technique, used by both AI Overviews and AI Mode, of issuing multiple related searches across subtopics and data sources to build a single response. It explains why these features often cite a wider and more diverse set of pages than a classic results page.
- Zero-click search: a search that ends without a click to any website. SGE made zero-click outcomes more common for informational queries, which is one reason AI visibility is now measured through mentions and citations rather than clicks alone.
Frequently asked questions
No. Google replaced the Search Labs experiment with AI Overviews in May 2024 and introduced AI Mode in 2025. The technology lives on, but Google no longer uses the SGE name. Articles that still describe "optimizing for SGE" are describing AI Overviews.
Google states there are no additional requirements and no special markup, AI text files or schema.org types needed to appear in AI Overviews or AI Mode. A page must be indexed and eligible for a snippet in classic Search. Standard SEO fundamentals apply: crawlable pages, clear text content, accurate structured data and helpful, reliable information.
Partly. AI features are part of Search, so blocking Googlebot in robots.txt removes a page from Search entirely. To limit what appears, Google points to the nosnippet, data-nosnippet, max-snippet and noindex controls. The Google-Extended token governs AI training and grounding in some of Google's other products, not Search itself.