AI Overviews: What they mean for brands and how to win visibility
AI Overviews are changing how people search. Here is what is really happening, what gets cited, and a practical framework to help your brand show up where it matters.
Why AI search changes visibility
AI Overviews synthesise information from several sources into a single answer at the top of the page. Users get what they asked for faster, which is good for them — but it means fewer clicks for brands that are not cited.
The important shift is subtle. Ranking well used to mean being findable. Increasingly it means being quotable: your page has to say something specific enough, clearly enough, that a model can lift it without distorting your meaning.
Our analysis of AI Overviews across high-value commercial queries shows a consistent pattern:
60%
of queries trigger an AI Overview
68%
of AI Overviews include cited sources
24%
fewer organic clicks on average
3.2
sources cited per AI Overview
What gets cited
Citations are not random. The sources that appear most often share a few traits.
Clear expertise: Named authors with real credentials, first-hand data and original research.
Answer-first content: Concise, well-structured answers placed near the question, not buried under preamble.
Strong entities: Brands, products and topics the model already understands and can disambiguate.
Topical depth: Comprehensive coverage that supports the sub-questions around the main one.
Technical trust: Fast, accessible, well-structured pages with clean markup and no crawl barriers.
A practical GEO framework
Use this five-step framework to improve your chances of being cited in AI Overviews.
Map intent and entities
Define the questions your audience actually asks and the entities the model needs to connect you to.
Create answer assets
Publish clear, expert-led answers that match real user questions, close to where the question is asked.
Prove trust
Add author bios, source data, case evidence and third-party validation.
Optimise the technicals
Ensure crawl access, fast performance and structured data that matches what is on the page.
Earn mentions
Build digital PR and partnerships that strengthen topical authority off your own domain.
How to measure progress
Track the right signals to understand impact and iterate quickly.
| Metric | What it tells you | How to track it |
|---|---|---|
| AI visibility | How often your brand is cited in AI Overviews | AI Overview tracking and brand mention analysis |
| Citation share | Your share of citations against competitors | AI Overview citation reports |
| Referrers | Traffic arriving from AI Overviews and assistants | GA4, Search Console, referrer analysis |
| Engagement | Whether the traffic you do win is worth having | GA4 engagement metrics |
| Conversions | Business impact from AI-influenced traffic | GA4 conversions and CRM data |
AI visibility trend
Share of tracked AI Overviews citing a client domain
| Period | AI visibility |
|---|---|
| Nov 23 | 12% |
| Dec 23 | 18% |
| Jan 24 | 27% |
| Feb 24 | 34% |
| Mar 24 | 46% |
| Apr 24 | 58% |
| May 24 | 68% |
Illustrative trend from one tracked programme. Citation share is volatile week to week — read the direction, not the individual point.
Common questions
Can I opt out of AI Overviews?
Partially. nosnippet, max-snippet and data-nosnippet limit what can be shown, but they also limit ordinary search snippets. For most commercial sites, removing yourself from the answer is worse than being one of the sources it is built from.
Does schema markup guarantee a citation?
No. Structured data helps a machine understand what a page is about and keeps your entity data consistent, which makes citation more likely. It is not a switch that turns citations on.
How often should we revisit this?
Quarterly for strategy, monthly for measurement. The underlying systems change frequently enough that an annual review will always be reacting to something that already happened.
Sources and methodology
The figures in this article are seeded sample values used to build and review the article template. Replace them with the results of the underlying XCL SEO study — including sample size, query set, geography, device and collection window — before publishing. Do not publish a statistic until its source is recorded.
Tom Rankin
Head of AI Search
Tom leads AI Search at XCL SEO. He helps brands earn visibility in AI Overviews through content, digital PR and technical excellence.
Expertise
- AI Search (GEO)
- Content Strategy
- Digital PR
- Technical SEO
- Analytics
Reviewed by Priya Raman, Technical SEO Director.
