From SGE to AI Overviews: What Changed

ClickRadius Institute · Published May 26, 2026

In three years, Google’s generative search went from a cautious opt-in experiment to the default way several billion people experience the web. The names changed along the way — SGE, then AI Overviews, now AI Mode as the front door — and each rename marked a real shift in how the system works, how often it appears, and what it takes to be visible inside it. If your mental model of Google’s AI search was formed in the SGE era (or even the early Overviews era), it is now materially wrong. This is the full arc, with the data.

Act I — SGE: the experiment (May 2023 – May 2024)

Google previewed the Search Generative Experience at I/O in May 2023: an opt-in Search Labs feature that generated an AI summary above the classic results, with source links attached. SGE was deliberately hedged — users had to enroll, coverage was narrow, answers were slow and verbose, and Google labeled the whole thing experimental.

But the experiment settled three questions that defined everything after:

• Format: the answer sits above the results, making the generated block — not position #1 — the top of the page;
• Sourcing: the answer cites a small set of sources, chosen by the system, not strictly the top-ranked pages;
• Intent: Google was willing to answer the query itself rather than merely referee links — a reversal of a 25-year-old contract.

The SGE year is when the discipline now called GEO got its academic foundation: the Princeton-led paper “GEO: Generative Engine Optimization” (released late 2023, presented at KDD 2024) tested what actually earns citations in generative answers and found that statistics, attributed quotations, and credible source citations measurably raised visibility — up to 40% in the authors’ reporting — while keyword stuffing did little.

Act II — AI Overviews: the rollout (May 2024 – early 2026)

At I/O 2024, SGE graduated: renamed AI Overviews and shipped to U.S. users by default, no Labs enrollment required. Expansion to more than 100 countries followed within months. The differences from SGE were substantive, not cosmetic:

• Selectivity. Overviews stopped trying to answer everything. They appeared where generation added value — informational, multi-step, and question-shaped queries — and stayed away from many navigational and transactional ones. Early rough edges (including some widely publicized wrong answers in the first weeks) pushed Google toward more conservative triggering and heavier grounding.
• Speed and compression. Answers got shorter, faster, and more confident, with citations presented in evolving formats — link cards, inline attributions, sidebars.
• Scale. Coverage grew steadily: by early 2026, industry trackers estimated Overviews appeared on roughly 15% of queries. Meanwhile Google began testing AI Mode — a fully conversational search tab — in Labs during 2025, expanding it through the year.

The click economics changed measurably in this act. Industry clickstream research put zero-click searches near 60% of all searches, up from an estimated ~45% before the generative era. Click-through studies found the #1 organic position’s CTR dropping from roughly 27% to roughly 11% on queries with an AI answer present. Ranking #1 under an Overview that resolved the query became the emblematic frustration of the period: technically winning a contest whose prize had moved.

Act III — AI Mode becomes the default (May 2026)

At Google I/O on May 19, 2026, the hedging ended. Google made AI Mode the default search experience globally — the conversational, Gemini-powered interface is now the front door, and the traditional ten blue links are the secondary view. Google’s own leadership framed the moment in unambiguous terms:

This is the biggest upgrade to our Search box in over 25 years.— Elizabeth Reid, VP of Search, Google (Google I/O 2026)

It’s our biggest upgrade to Search ever.— Sundar Pichai, CEO, Google (Google I/O 2026)

The numbers moved with the default switch. AI Overviews now appear on roughly 48% of queries — up from about 15% in early 2026, per industry tracking. And within AI Mode itself, behavior is starkly answer-native: roughly 93% of AI Mode sessions end with zero clicks to any website. Google also introduced Information Agents for AI Pro and Ultra subscribers — autonomous agents that monitor topics continuously, run their own searches, and deliver summaries to the user without a single site visit.

Read plainly: Google is no longer a referral engine with an AI feature. It is an answer engine with a link archive attached — and it refers users outward mainly when a source offers genuine expertise or authority the AI cannot replicate on its own.

What actually changed, era by era

What stayed the same — the through-line for businesses

Across all three acts, the selection logic has rewarded the same fundamentals, which is the most strategically useful fact in this whole story:

1. Retrievability. Every era pulled candidates from Google’s index. Blocked crawlers, rendering failures, and thin indexation have been disqualifying since day one.
2. Extractability. From SGE to AI Mode, systems favored content they could lift cleanly: direct answers under question-shaped headings, structured data, self-contained sections.
3. Evidence. The KDD 2024 findings — statistics, quotations, citations — have only grown more relevant as grounding got stricter. Engines protect their own credibility by citing sources that carry verifiable substance.
4. Entity trust. Industry data suggests the majority of what drives AI citations is off-site: consistent entity data, directory presence, third-party corroboration. AI Mode’s deeper reasoning leans harder on knowing who you are, not just what the page says.

What changed is the stakes. Under SGE, ignoring generative search cost you nothing. Under default AI Mode with ~48% AI-answer coverage and Information Agents browsing on users’ behalf, being uncitable means being absent from roughly half of Google — and from every agent-delivered summary. Yet industry data still shows a large majority of brands have zero AI-search mentions. The penalty for absence rose; the crowd competing for presence has not caught up.

How each era rewrote the practitioner’s job

The three acts did not just change Google’s interface; each one quietly rewrote what search work is. Tracing that evolution explains why teams that adapted act-by-act are so far ahead of teams now confronting all three changes at once.

The SGE era created the observer’s job

During the experiment, the rational move was instrumentation: build prompt sets, watch which sources the generated answers drew on, and notice — earlier than competitors — that citation selection did not simply mirror rankings. Teams that started measuring in 2023–24 entered the Overviews era with baselines; everyone else entered it with opinions.

The Overviews era created the dual-scoreboard job

From May 2024 onward, rankings and citations had to be tracked side by side, because they began to diverge on real revenue queries. This is also when the evidence triad from the KDD 2024 research moved from academic finding to content standard: pages rewritten answer-first with statistics, quotations, and sources started appearing in Overview citations that their keyword-optimized predecessors never earned.

The AI Mode era creates the entity manager’s job

A conversational default interface, answering follow-ups and dispatching Information Agents, leans on knowing who is trustworthy — which makes the off-site entity footprint (directories, corroborating coverage, consistent data) a first-class responsibility rather than an afterthought. The practitioner’s center of gravity has moved, in three acts, from the page to the answer to the entity.

The compounding lesson across all three acts: Google telegraphed every step roughly a year before it became default behavior. SGE previewed Overviews; Labs-era AI Mode previewed the 2026 default. Which means the current previews — agents that search on the user’s behalf — are not speculation about the next act. They are the rehearsal happening in public.

The tactical checklist for the AI Mode era

• Audit robots.txt and rendering for Google’s crawlers — including Google-Extended policy decisions made years ago and forgotten.
• Restructure priority pages answer-first, one canonical answer per question, with Article/FAQPage/Organization JSON-LD.
• Enrich those pages with the evidence triad: real statistics, attributed quotes, credible external sources.
• Reconcile your entity across the web — name, category, location, facts — until every source a triangulating engine checks agrees.
• Monitor citations with real prompts, on a schedule, across engines — not just Google. ClickRadius tracks five live AI engines (ChatGPT, Gemini, Perplexity, Claude, Grok) and scores site readiness across six categories, because AI Mode’s behavior shifts with every model update and last quarter’s screenshot is not a strategy.

Frequently asked questions

Is SGE the same thing as AI Overviews?

SGE (Search Generative Experience) was the experimental, opt-in version Google ran in Search Labs starting in May 2023. AI Overviews is what it became when Google shipped it to everyone beginning in May 2024 — refined, faster, more selective about when it appears, and backed by Gemini. Any article discussing SGE tactics is describing the prototype era; the principles carried forward, but coverage, trigger behavior, and citation formats have all evolved substantially since.

What is the difference between AI Overviews and AI Mode?

AI Overviews is a generated answer block placed on the classic results page, above the familiar links. AI Mode is a fully conversational search experience — a chat-style interface powered by Gemini that handles follow-up questions and complex tasks, with links and citations woven into its responses. At Google I/O 2026 the company made AI Mode the default search experience globally, which moved the traditional ten blue links to secondary status rather than eliminating them.

How do I get my business cited now that AI Mode is the default?

The levers are consistent across the whole SGE-to-AI-Mode arc: be cleanly indexable to Google’s crawlers, answer real questions directly in extractable structures, carry the evidence signals research links to citation (statistics, attributed quotations, credible sources), and build a consistent entity footprint across the web — industry data suggests off-site signals drive the majority of AI citations. Then verify with real prompts on a schedule, because citation behavior shifts with every model update.

Three eras in, the question is the same: can the engine cite you? Get your free AI Readiness Score to find out in minutes, or see ClickRadius pricing for continuous monitoring across five AI engines.