Google AI Mode and AI Overviews don’t cite the same sources: measure them separately

Are AI Mode and AI Overviews the same product?

No. AI Overviews are a summary shown at the top of some search results. AI Mode is a separate conversational experience built for complex questions and follow-ups.

Google documents both on one page and applies the same entry rules. A page that is indexed and eligible for a snippet can be cited, and Google says there are no additional technical requirements ^[1].

The same page carries a line most brands skip. Google says the two “may use different models and techniques,” so the answers and links they show will vary ^[1].

Both use query fan-out: Google splits a question into subtopics and runs multiple related searches before writing the answer ^[1]. AI Mode takes this further, with longer answers and follow-up questions.

Scale is no longer a reason to ignore it. Google reported that AI Mode passed 1 billion monthly active users after its global expansion, in its Q2 2026 remarks ^[2].

Do AI Mode and AI Overviews cite the same sources?

Very little. Two independent studies, on different samples, put URL overlap between 10% and 14% for the same query.

• Ahrefs, US data from September 2025. Across 540,000 query pairs analyzed for citations, 13.7% of cited URLs were shared by both surfaces. For the top three citations, overlap reached 16.3% ^[3].
• SE Ranking, 10,000 keywords analyzed in June 2025. 10.7% of URLs and 16% of domains were shared ^[4].

Domain overlap runs higher than URL overlap. Google often draws on the same websites, but not the same pages. A comparison page cited in AI Overviews guarantees nothing in AI Mode, even when the rest of your site is well known to Google.

The meaning of the answers converges, though. Ahrefs measured 86% average semantic similarity, while shared wording stayed at 16% ^[3]. The two surfaces often say the same thing, with different words and different evidence.

What changes for a B2B brand?

AI Mode leaves more room for brands. AI Overviews name fewer of them and favor different types of sources.

Ahrefs: US data, September 2025. SE Ranking: 10,000 keywords, June 2025.

In 61% of cases, AI Mode includes every entity the AI Overview named, then adds more ^[3]. AI Mode is often the long list and the AI Overview the short one. The short list is harder to get into, and worth more on queries where a summary appears.

Sources carry different weight too. Quora shows up 3.5x more often in AI Mode citations, while Reddit is cited at a similar rate on both ^[3]. A content plan built for one surface leaves gaps on the other.

Can Google Search Console separate the two?

No. Google counts appearances in AI features as part of overall Search Console traffic, under the Web search type ^[1].

Search Console tells you whether organic clicks are rising or falling. It does not tell you whether your brand was named in an AI Overview, in AI Mode, in both or in neither. Nor does it show which competitors were named in your place.

Variance makes a single check even less reliable. Tracking more than 43,000 keywords for a month, Ahrefs found that 45.5% of the sources cited by an AI Overview are replaced from one generation to the next ^[5]. SE Ranking ran the same queries three times in one day in AI Mode, and only 9.2% of URLs appeared in all three runs ^[4].

A screenshot is not a measurement. Repeat each question on each surface, and compare series.

Google merged the two experiences. Should you wait?

No. The overlap figures come from 2025 data. They need to be rechecked on your own buying questions, not dismissed.

In July 2026, Google said it had brought AI Overviews and AI Mode together into “one seamless Search experience” ^[2]. That describes how users move between them. It does not say the two now cite the same pages, and Google’s documentation still says they may use different models and techniques ^[1].

The Ahrefs and SE Ranking studies used US data from 2025. Nobody knows the real overlap for your category, on your buying questions, today, until it has been measured.

How do you get cited on both surfaces?

The fundamentals are shared. The priorities are not.

• Shared baseline. Pages that are indexed and eligible for a snippet, with no nosnippet rule on the content you want reused. That is Google’s only stated condition for both ^[1].
• For AI Overviews. Short, direct answers at the top of the page, and the formats this surface cites more often, including video ^[3].
• For AI Mode. Cover the sub-questions fan-out will generate: selection criteria, comparisons, use cases, integrations. Keep reference third-party sources accurate too; Wikipedia is the clearest gap in Ahrefs’ data ^[3].
• Track each surface as its own series. Several runs per question, before and after each publication, with the same protocol.

How does ZivRank measure these surfaces?

ZivRank measures how ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews present B2B brands on the buying questions of their market. Google AI Overviews searches are reported as a separate series and never added to assistant answers. Each question runs 5 times, and a ±15% noise threshold is set before any comparison. The full protocol is on our methodology page.

AI Mode is not part of the standard protocol. The rule still holds: a surface that cites different sources is measured on its own, never inferred from another. The goal is to win the AI shortlist your buyers see, by finding the sources and citations that matter and producing the content that is missing. ZivRank commits to the work, the measurement and the production, not to a guaranteed citation.

To see where your brand stands today, the free Quick Scan tests up to 5 of your buying questions on ChatGPT and Gemini, with an expert video walkthrough within 48 hours. Full programs are on the pricing page.