How to Boost Your Brand's Credibility with Google's New Citation Count Feature

What did Google change in AI Mode on May 6, 2026?

On May 6, 2026, Google rolled out what it described as the most significant structural update to AI Overviews and AI Mode since those features launched. Five changes arrived together. Two are architectural. Citations no longer collect in a pill stack at the bottom of an AI-generated answer. They now attach inline, anchored to the specific sentence or claim they support. On desktop, hovering over any inline link surfaces a preview card showing the source's site name and page title before the user clicks through.

The third change is the one every brand should care about: Google introduced a 'highly cited' label. Sources that appear frequently across AI answers on a given topic can now earn that badge in the interface. Citation frequency, which had been invisible infrastructure inside Google's ranking and retrieval systems, became a visible trust signal to ordinary searchers. A brand tagged as 'highly cited' in an AI answer looks categorically different from a brand that is referenced once and unlabeled.

Two further changes arrived in the same update: a preferred sources feature letting users weight publications they subscribe to or already trust, and a perspectives carousel surfacing first-hand voices from forums and review sites. Together, the five changes shift the search result from a list of links to a curated editorial package, with citation count acting as the authority marker at the top. Source: Search Engine Land, 'Google updates links within AI Overviews and AI Mode,' May 2026.

What does the 'highly cited' label actually signal to searchers?

Before May 6, a brand could be cited in hundreds of AI answers without a searcher ever knowing it. Citation count was a signal Google used internally to assess source reliability, but it was invisible in the interface. That changed with this update. When Google attaches a 'highly cited' label to a source, it tells the user plainly: this brand has been the answer enough times that we treat it as authoritative on this topic. It is the first time Google has made that distinction public.

The credibility implication compounds across sessions. A user who sees your brand labeled 'highly cited' in a software comparison query, then in a procurement query, then in a local service query in the same week builds a mental picture of your brand as a category authority. That is a function traditional blue-link search never performed. It showed you where you ranked, not how often you were right. The 'highly cited' label is the first time Google has surfaced citation frequency as consumer-facing proof.

For competitive categories, this creates a visibility flywheel. Brands already appearing across many AI answers pick up the label. The label increases click share within AI Mode. Higher click share sends behavioral signals back to Google. Those signals reinforce further citation. Brands not yet inside that loop need to enter it before competitors lock up the label on the queries that matter most to their category. The window is not infinite: citation share compounds in the same way domain authority compounded in traditional SEO, and the leaders are already ahead.

Who is most affected by this change?

The brands hit hardest are those that built their digital presence almost entirely on organic search rankings. Until recently, being in the organic top 10 for a keyword provided reasonable assurance of appearing in AI-generated answers on that topic. That assurance is eroding fast. Top-10 organic rankings explained approximately 76% of AI Overview citations in mid-2025. By early 2026, that figure had dropped to roughly 38%, according to analysis aggregated across multiple query-set audits. A ranking that once implied citation now frequently does not.

B2B SaaS and tech growth teams are especially exposed. Their category decisions (best [tool], [tool] vs [competitor], top [tool] for [use case]) are exactly the kinds of queries where AI Mode now synthesizes an answer instead of returning a list of links. If your brand does not earn a citation in those synthesized answers, you are not on page two. There is no page two in an AI answer. You are simply absent from the conversation where your buyer made their decision.

Service businesses competing on local queries face the same pressure in a different format. When someone asks an AI which plumber, accountant, or agency to hire in their city, the AI cites a small cluster of sources it trusts on that topic. Citation share on 'service] in [city]' prompts is the new measure of local visibility, and the 'highly cited' label in AI Mode is now the consumer-facing proof that you are the authority. Brands investing in [generative engine optimization ahead of this update are already ahead.

Why do brand mentions predict citation count more than backlinks?

The Ahrefs 75,000-brand study, published in May 2026, is the most comprehensive analysis to date of what actually predicts AI citation visibility. Its central finding is striking: brand web mentions correlate 0.664 with AI Overview citation visibility. Backlinks correlate 0.218. The signal that has driven search engine optimization for two decades is roughly one-third as predictive of AI citation as the signal that PR programs and earned media produce. YouTube mentions scored highest of all signals studied: a Spearman correlation of 0.737, defined as any time a brand name appears in a video title, transcript, or description.

The mechanism is not mysterious once you consider how large language models are trained. AI systems learn what is true and authoritative from the web they were trained on. Brands that appear in authoritative third-party contexts, that are discussed and cited rather than just linked to, encode into those models as reliable sources on their topics. A backlink from a high-authority domain tells a crawler your page matters. A mention in a well-regarded editorial tells a model your brand knows what it is talking about. Those are different signals producing different outcomes.

This matters practically because the work required to move each metric is different. Building backlinks means producing content that other sites link to. Building brand mentions means producing content that other sources refer to, cite, and quote: real data, original research, positions your brand takes publicly, comparisons with attributed sources. That is exactly the kind of content AI models prefer to surface. The May 2026 update, with its visible 'highly cited' label, rewards brands that have already made that shift. For brands that have not, it makes the cost of not making it visible for the first time.

How should your brand respond to the new citation count feature?

The first move is an honest audit of your current citation share across all AI surfaces, not just Google. AI Mode and AI Overviews cite the same URLs only 13.7% of the time, which means visibility on one surface does not transfer to another. A brand that earns frequent citations in ChatGPT answers on its category queries may be completely absent from Google AI Mode answers on the same questions. Measure citation share per engine, per query cluster, and per competitor so you know where you actually stand before the 'highly cited' label goes to someone else.

The second move is to build a content program structured around citable assets: original data points, comparison frameworks, expert sourcing, and answer-first structures that give AI models a specific, attributable claim to extract. An article that states a position with a named source, a dated statistic, or a reproducible comparison result is far more likely to earn a citation than an article that summarizes a category in general terms. General content fills a site. Citable content earns a 'highly cited' label. The distinction shows in the Ahrefs data and it now shows in the Google interface.

The third move is to build a systematic pipeline for third-party mentions. That means earning coverage in publications AI engines retrieve, appearing in podcast transcripts where your brand name gets mentioned alongside your category, and earning placement in comparison and review content produced by sites your buyers already trust. The Ahrefs data makes the case: off-site brand signals are now the primary driver of AI citation, not on-site content volume or domain authority in the traditional link sense. The May 2026 citation count feature made that gap visible to every searcher. The brands that respond now will hold the 'highly cited' label when it matters.

1. Audit citation share across ChatGPT, Perplexity, Gemini, Copilot, and Google AI Mode separately: the overlap between surfaces is only 13.7%
2. Identify the 10 to 20 queries where your category decision is made inside an AI answer and measure your citation count per query cluster
3. Create at least one citable asset per topic cluster: a real dated statistic, an original comparison, or a data study with a named source
4. Build a systematic earned-media program targeting publications and channels that AI engines cite frequently in your category
5. Track 'highly cited' label status in Google AI Mode alongside citation share month over month, not just keyword rankings

Sources

Google rolled out five structural changes to AI Overviews and AI Mode on May 6, 2026, including inline citations, hover preview cards, and a 'highly cited' label Search Engine Land, 2026-05-06

Google AI Overviews and AI Mode gained preferred sources, a perspectives carousel, and highly cited labels as part of the May 2026 update Search Engine Land, 2026-05-06

Brand web mentions correlate 0.664 with AI Overview citation visibility versus 0.218 for backlinks, across 75,000 brands; YouTube mentions score 0.737 Ahrefs, 2026-05-26

YouTube mentions are the strongest signal of AI visibility across 75,000 brands studied, per Ahrefs research covering 730,000 AI responses Business Wire, 2026-05-26

Google I/O 2026 search updates including AI Mode expansion and new citation surfaces Google Blog, 2026-05-20