On May 27, 2026, Google rolled out three changes to its AI search interfaces that collectively shift how authoritative sources get identified, displayed, and clicked. Most coverage treated these as UI updates. They’re not — they’re structural changes to how citation authority is surfaced to users, and they have direct implications for any brand trying to appear in AI Overviews or AI Mode.
The three features: a “Highly Cited” label expansion across web results, a “Preferred Sources” label in AI Mode, and a Perspectives carousel that prioritizes firsthand accounts and forum content. Each one changes the visibility calculus for AI search differently. Here’s what each means and what to do about it.
What the “Highly Cited” Label Actually Signals
Google’s Highly Cited label marks articles that are “explicitly referenced by other articles” — i.e., primary reporting that downstream content cites back to. The label is now expanding across more web result links in Google Search generally, not limited to AI experiences.
This matters for AI visibility for one specific reason: AI Overviews and AI Mode pull from the same source authority graph that Google uses to determine what earns a Highly Cited designation. If your content is original primary reporting — original data, first-party survey results, original analysis — it has a higher probability of being cited by other articles, which increases the probability of earning the Highly Cited label, which correlates with being pulled into AI-generated answers.
The practical implication: content that gets cited by other publishers is increasingly tagged as structurally more authoritative than content that merely ranks well. A page at position 8 with three inbound citations from industry publications may outperform a page at position 2 with zero citations in terms of AI inclusion. This is the same dynamic the Opollo 2026 benchmark surfaced when it found AI visitors convert at 14.2% vs. Google organic’s 2.8% — AI search is pulling later-funnel, more-evaluated content.
Preferred Sources: 345,000 Sources Selected, 2× Click Rate
The Preferred Sources feature lets users designate specific domains they want to see prioritized in AI responses. Google reports that users have selected over 345,000 unique sources globally, and users are twice as likely to click through to preferred sources compared to other AI-cited links.
This creates a flywheel that doesn’t exist in traditional SEO. If your domain is a user’s preferred source, your citations in AI Mode generate 2× the click-through rate. But to become a preferred source in the first place, you have to be cited enough that users discover you through AI answers, trust the content, and actively add you. The brands that build AI visibility first are the ones most likely to accumulate preferred source designations — compounding their advantage over time.
For B2B brands, this makes the category authority play even more critical. If you’re the domain that consistently provides the best data on your category — the original survey, the benchmark report, the methodology — your domain becomes a natural preferred source candidate for professionals researching that space. The 345,000 figure is still small relative to total users; it represents early adopters who are already optimizing their AI search experience, which skews toward high-value B2B buyers.
The Perspectives Carousel and What It Means for Forum Visibility
The new Perspectives carousel surfaces “helpful perspectives from online discussions, forums, and social media” on developing topics. Combined with Google’s existing integration of Reddit content into AI Overviews (announced in the May 2026 update), this represents a structural elevation of community-generated content in AI search results.
Research from Discovered Labs documented that Reddit appeared in only 0.35% of visible ChatGPT citations, but occupied 27% of ChatGPT’s internal search slots during processing — a massive gap between what’s shown and what influences the AI’s reasoning. Google’s Perspectives carousel is making the community influence more explicit and visible.
The tactical response: if your brand or product category has significant discussion on Reddit, Quora, or niche forums, you need to be present in those conversations in a way that reflects accurate, current information about your product. Inaccurate or outdated community information about your brand will influence AI answers even if it never shows up in a traditional citation. Monitor and, where appropriate, participate.
FAQ Content: The Overlooked On-Ramp to AI Overviews
Semrush’s analysis of AI Overview queries found that more than 80% are informational, and 82% have average monthly search volumes under 1,000. This is a fundamentally different distribution than traditional SEO targets — AI Overviews are winning queries that most brands never optimized for because the traffic volume didn’t justify it.
Five content sources that consistently produce FAQ content that earns AI Overview inclusion:
- Google Search Console question queries — Filter for queries with 8+ words and positions 4–20. These are long-tail informational searches where you rank but don’t get clicks. An FAQ page targeting these queries precisely often gets pulled into AI Overviews faster than polished pillar content because it directly matches the query structure.
- People Also Ask clusters — PAA boxes reveal how Google groups related questions. Tools like AlsoAsked and AnswerThePublic surface question trees at scale. Build FAQ sections that address the full cluster, not just the head question.
- Customer support logs — Your support team receives the real questions your customers ask in their own language. These questions, formatted as FAQ entries with direct answers, consistently outperform marketing-written content for AI inclusion because they match actual user phrasing.
- Reddit and community forums — Sort by Top within your product category. The questions at the top of relevant threads represent high-demand informational needs with authentic phrasing that AI systems recognize as natural query language.
- AI prompt data tools — Emerging platforms like Writesonic now surface aggregated AI prompt data, revealing the conversational queries users ask AI tools before turning to traditional search. These are your pre-search queries — the questions that exist in the AI layer before they ever hit Google.
The “Highly Cited” Execution Playbook
To build content that earns Highly Cited status over time, three content types have the highest citation rates from other publishers:
Original primary data. If you run a survey, publish the full methodology and dataset. Other publishers cite primary sources, not summaries. A study with n=200 and a clear methodology will be cited more than a 2,000-word synthesis piece. The AuthorityTech analysis found that brand mentions across third-party sources correlate 3× more strongly with AI visibility than backlinks — and Highly Cited content generates exactly this kind of third-party mention pattern.
Category benchmark reports. Annual or quarterly benchmark data for your industry becomes a citation magnet because other content creators need something to cite when making claims about the category. If you publish the benchmark, you own the cited source position. Multiple GEO reports have now confirmed that Perplexity cites 8.2 sources per answer — any brand that publishes original category data in a citable format has a structural advantage in that citation pool.
Definitional or methodology content. When your content establishes the accepted definition of a concept or methodology in your category, it becomes the default citation for anyone writing about that concept. This is high-effort, long-horizon work, but it compounds in ways that no other content type does — including in AI training data and retrieval weighting.
Schema Markup: Still the Most Underused Lever
Schema coverage delivers 3–5× more AI citations according to AuthorityTech’s analysis, yet 80% of B2B websites have incomplete or missing schema implementation. The specific schemas most relevant to AI Overview inclusion in 2026:
- FAQPage schema — Marks up Q&A content in a format AI retrieval systems can parse structurally, not just semantically.
- Article schema with author and datePublished — Freshness signals matter. AI systems show preference for recently published content in rapidly evolving categories. A page with accurate datePublished schema is easier for AI retrieval to evaluate for freshness than one without.
- Organization schema with sameAs links — Cross-referencing your brand’s Wikipedia page, LinkedIn profile, and other authoritative profiles helps establish entity disambiguation. AI systems that can confidently identify which entity you are cite you more consistently.
- HowTo and Dataset schemas — For instructional and data content, these schemas signal structural authority that correlates with the content types AI Overviews preferentially cite.
Putting It Together: The Three Tracks You Need to Run Simultaneously
The Discovered Labs analysis documented the core challenge for 2026 GEO strategy: the correlation between Google rankings and AI citations dropped from 76% to 38% in roughly twelve months. That means traditional SEO, AI citation optimization, and agent-source-trust are now three distinct optimization tracks that only partially overlap.
The new Google features make the citation track more visible to users — Highly Cited labels, Preferred Sources, Perspectives carousels — but they don’t make it automatic. Earning Highly Cited status requires producing citable primary content. Becoming a Preferred Source requires consistent AI citation presence that builds user trust over time. Appearing in Perspectives carousels requires authentic participation in the communities AI systems are increasingly indexing.
None of this requires a complete marketing overhaul. It requires redirecting a portion of content production toward primary data, redirecting a portion of distribution toward community participation, and implementing schema markup that most sites still haven’t deployed correctly.
Check Where You Stand Before You Strategize
Before deciding which of these tracks to prioritize, you need to know your current AI visibility score. Most brands operating in competitive B2B categories score under 40/100 on AI visibility audits — meaning AI systems frequently omit them from answers where they should appear, even when they rank well in traditional search.
Run a free AI visibility audit at ai-visibility.llmagnet.com to see how ChatGPT, Perplexity, and Google AI Overviews currently represent your brand and category. The audit surfaces specific gaps — whether the issue is entity disambiguation, schema coverage, or citation footprint — so you can prioritize the right track rather than spreading effort across all three simultaneously.
The Highly Cited label is now visible to users. The Preferred Sources feature is compounding click rates for cited brands. The Perspectives carousel is amplifying community voices. The brands that start building citation authority now are the ones who will see the 2× click multiplier and the compounding Preferred Source flywheel — before those advantages become as competitive as traditional search rankings already are.