Google AI Overviews in 2026: How SEO Is Changing

In May 2024, Google launched AI Overviews — a block of generative answers at the top of search results, assembled from multiple sources and resolving a significant portion of a query directly in the search engine, without the user visiting any website. By 2026, AI Overviews appear in roughly 18–22% of commercial and informational queries in English, and are gradually expanding to Russian-language searches through Google.ru.

This changes the rules of the game — and not always for the worse. In this article we break down what AI Overviews are technically, what they do to traffic, and what practical steps an SEO specialist should take in 2026.

What AI Overviews Are

This is a block at the top of Google's search results, occupying space above the standard ten organic links. Inside the block is an answer to the user's query generated by the Gemini neural network, with citations from 3–8 sources (displayed as small cards on the right or below).

Here is how it works technically:

The user enters a query
Google decides whether the query is suitable for an AI Overview (complex informational queries — yes; narrow transactional ones — usually no)
The system selects sources for generation from the classic top-10 results
Gemini synthesises an answer, not copying the source text verbatim but paraphrasing it
The block is displayed above the results with link cards pointing to the sources

The key detail: you can only appear in an AI Overview from within the top results. If your site is in position 15, it is guaranteed not to be cited. If it is in the top 3, the odds are high — but not guaranteed.

What This Does to Traffic

The worst news for informational sites: AI Overviews noticeably reduce CTR.

According to analyst data (SEMrush, Ahrefs, and SEO community surveys from 2024–2026), the overall picture looks like this:

Query type	CTR without AI Overview	CTR with AI Overview
Informational ("what is X")	28–35% at position 1	12–18% at position 1
Comparative ("X vs Y")	32–40% at position 1	14–20% at position 1
Commercial ("buy X in Moscow")	28–35% at position 1	AI Overview rarely appears
Branded ("company Z website")	70–85% at position 1	AI Overview rarely appears

The pattern: on purely informational queries, some users are satisfied by the answer in the AI Overview and do not click through. On commercial and branded queries, AI Overviews appear infrequently, and CTR there has barely suffered.

The Paradox: Being in AI Overview Is Better Than Ranking #1

A site cited in an AI Overview receives significantly more brand recognition and trust than a site in the ordinary position one. Based on early research from 2024–2025:

Roughly 8–12% of users click on the source cards within an AI Overview
But brand awareness generated by an AI Overview mention grows considerably faster than through a standard snippet
Branded queries (where a user later searches specifically for your company by name) grow by 30–45% after AI Overview citations

In other words, traffic may dip in the short term, while in the long term audience quality and brand recognition improve.

What Determines Inclusion in an AI Overview

Google does not publish exact ranking factors, but from practice and observation of real AI Overview sources, the following is clear.

1. Schema.org Structured Markup

The most frequently cited sites carry markup for:

Article with author and date specified
FAQPage — especially for question-and-answer content
HowTo for step-by-step instructions
Product for commercial reviews
MedicalCondition, Recipe, Event — for specialised niches

Without structured markup, appearing in AI Overviews is significantly harder — Gemini relies on structured data when selecting sources.

2. Thesis-Driven Content Format

AI Overviews are short answers. Content that is easier to "slice" into short fragments for citation gets cited more often.

In practice this means:

A direct answer in the first paragraph. If the article is "What is X", the first sentence must be a definition. Not "in this article we will explore…", but "X is [definition]."
Lists and tables are easier to structure than continuous prose. If you are making a comparison — use a table. If you are listing items — use a list.
Short paragraphs of 2–3 sentences. Gemini compresses or ignores long walls of text.
Subheadings phrased as questions. "How much does X cost?", "How does X differ from Y?" — the search engine can easily cite the corresponding block.

3. E-E-A-T Signals

The same signals that matter for classic SEO in YMYL niches also matter for AI Overviews — even in ordinary niches. Named authors, update dates, domain authority, and citations in other sources. Google needs to "trust" the source before placing it in the AI block.

4. Content Freshness

Significantly outdated content rarely appears in AI Overviews. For topics that change over time (technology, markets, legislation), an article from three years ago is almost guaranteed not to be cited. Regular updates — at least once a year — with a review of the facts and an updated modification date are essential.

5. Unique Information Not Found in Typical Sources

Paradoxically, AI Overviews prefer to cite sites that provide unique information — original data, research, and expert opinions that are absent from typical top-ranking content. If your content is a retelling of Wikipedia, your chances of appearing in an AI Overview are minimal.

What to Do Right Now

A practical checklist for adapting an existing site to the AI Overviews era:

Audit your existing content. What are the 10–20 highest-volume queries your site answers? Which of them already show AI Overviews in Google? (You can check manually or via tools such as SERPRobot or AccuRanker.)
Rewrite your opening paragraphs. In every article, the first sentence must be a direct answer to the question the article addresses. No "in this article we will", no watery introductions.
Add FAQ blocks. On key topic pages — add a dedicated block with 5–8 typical questions and concise answers, marked up with the FAQPage schema.
Restructure your subheadings. Rewrite H2 and H3 headings as questions where this is natural for the topic. "Advantages of X" → "What are the advantages of X."
Add structured markup. Article, Author, datePublished, dateModified on every article. This is basic hygiene in 2026.
Update old posts. All content more than a year old — review it, update the dates, add new facts.
Remove duplicative filler content. 800-word articles that say nothing new will not bring traffic in the AI Overviews era. It is better to have 30 in-depth pieces than 200 superficial ones.

What to Do About a Drop in Informational Traffic

If your site depended on organic traffic from "what is X" queries, a strategic adjustment is needed.

Informational traffic does not disappear entirely, but it does decline. To compensate:

Shift your content focus toward commercial and comparative topics. "What is a CRM" → "10 CRM Systems for Small Business in 2026 — a Comparison." AI Overviews do appear on comparative queries, but users click source cards more actively there.
Build branded traffic. Brand recognition is something AI Overviews cannot take away. Content marketing, expert publications, YouTube — everything that builds a "brand of expertise."
Diversify beyond Google. Yandex Zen, VK, Telegram channels, industry platforms. Traffic diversification is a defence against any algorithmic fluctuation.

Summary

AI Overviews are not "the end of SEO", as many proclaimed in the first months after launch. They are a rule change: fewer clicks on informational queries, but greater brand recognition and better-quality traffic on commercial ones. Sites that adapt to the new citation logic will win — those that keep writing "SEO texts of 800 words with keyword density" will gradually drop out of search entirely.

Want to find out whether your site appears in AI Overviews for your key queries, and what to do to appear there more often — get in touch.