What Was Search Generative Experience (SGE)?

Search Generative Experience (SGE) was Google's first public generative-search product: an opt-in experiment, announced at Google I/O on May 10, 2023, that placed an AI-generated "snapshot" above traditional results for users enrolled in Search Labs. It was the direct precursor to AI Overviews, which replaced it as a default feature in May 2024.

What SGE actually was

SGE generated a multi-sentence summary grounded in Google's Search index, displayed source cards alongside the text, and offered follow-up questions that opened a conversational view. It ran on the PaLM 2 model family at launch, later moving to Gemini models. Because it was opt-in and US-first (with expansions to India and Japan in 2023), its user base skewed toward early adopters — a key caveat for every SGE-era study.

What changed between SGE and AI Overviews

The strategic shift mattered more than the rename: an experiment became the default reading experience for billions of searches, moving generative answers from a curiosity to the primary surface brands compete on.

Why the history is useful to practitioners

Early SGE studies measured dramatic click loss and near-total answer coverage — findings that partly reflected the experiment's aggressive trigger settings rather than durable behavior. When Google shipped AI Overviews, trigger rates initially dropped, then climbed again through 2025. Reading visibility research without checking which era it measured is a common analytical error; the what-is-GEO guide covers how to benchmark against the current surface instead.

Example

A 2023 blog post claiming "84% of queries show generative results" describes SGE Labs behavior, not today's production reality — Pew measured 18% AI-summary incidence on Google in March 2025. Treat SGE numbers as historical context, and track the live surfaces defined throughout this glossary.