Cited — 2026-05-02

Welcome to this week's edition of the Cited newsletter, where we track what is actually shifting in AI search visibility so you do not have to spend hours piecing it together yourself. This issue covers the developments that landed hardest between late April and the first week of May 2026. If you are running marketing at a growing business, these are the signals worth paying attention to.

What Changed in AI Search This Week

Google AI Overviews Are Pulling From Deeper in the Index

One of the more significant shifts we noticed this week: Google AI Overviews are increasingly sourcing content from pages that do not rank in the traditional top 10 organic results. In several verticals including B2B SaaS, financial services, and professional consulting, pages sitting at positions 15 to 30 in organic search are being cited in AI-generated answers.

This matters because it decouples AI visibility from traditional SEO rank, at least partially. A page that never broke page one can still appear prominently in an AI Overview if the content is structured correctly and the entity signals are strong. We are tracking this across client accounts and the pattern is consistent enough to act on.

What to do: audit the pages on your site that cover specific questions well but rank poorly. Those are your first candidates for structured content optimisation targeting AI extraction.

ChatGPT Source Diversity Is Narrowing

A counterintuitive finding this week. Despite ChatGPT expanding its browsing capabilities throughout Q1 2026, the actual diversity of sources it cites in responses appears to be narrowing rather than broadening. Our internal tracking data across 40+ brands shows that ChatGPT is increasingly anchoring on a smaller set of high-authority sources per topic cluster.

The implication is clear: being one of three or four trusted sources for a topic is more valuable than appearing occasionally across a wider range of queries. Depth of topical authority is pulling ahead of breadth.

Perplexity's "Pro Search" Mode Is Behaving Differently

Perplexity's Pro Search mode, used by its paid subscribers, is showing meaningfully different citation patterns compared to standard queries. Pro Search pulls more sources, weights recency more heavily, and appears to favour content with explicit data points over general explanatory content.

If your audience skews toward power users and researchers (common in B2B), optimising for Pro Search behaviour is worth prioritising. That means publishing with dates, updating content frequently, and embedding specific statistics rather than general claims.

The Bigger Picture: What These Signals Mean for Your Programme

The Case for Topical Depth Over Keyword Volume

For years, content strategy was driven by keyword volume. More searches meant more traffic potential, so that is where the effort went. AI search is disrupting that logic. The models selecting which sources to cite are not optimising for the most-searched phrase. They are selecting sources that demonstrate the clearest, most complete understanding of a topic.

A business that publishes ten thin articles across ten loosely related topics is unlikely to become a trusted AI source for any of them. A business that publishes three genuinely comprehensive resources on a narrow topic cluster, updated regularly, has a much better shot. This is not a new principle in SEO, but AI search is making it non-negotiable.

Counterargument Worth Considering

Some teams push back on topical depth strategies because the content effort is significant and the attribution is harder to measure than traditional rank tracking. That is a fair concern. AI visibility does not show up cleanly in Google Analytics. Citations in ChatGPT do not always generate direct clicks. The business case requires a different measurement framework.

The honest answer is that AI-driven citation traffic is real but currently harder to isolate. Research from BrightEdge published in early 2026 found that AI-sourced traffic carries higher engagement metrics on average, including longer session times and lower bounce rates, suggesting intent alignment is strong even if volume reporting is imperfect. The infrastructure for measuring this properly is still catching up to the behaviour.

Schema, Entity Clarity, and Why Both Still Matter

Two things we keep seeing in the data: businesses that have clean structured data markup continue to outperform those that do not, and businesses with strong entity disambiguation (clear signals about who they are, what they do, and for whom) are being cited more consistently across models.

Entity clarity is not just about schema. It includes how consistently your brand, people, and products are described across your own site, your social profiles, third-party coverage, and citations elsewhere on the web. The more coherent that picture, the more confidently an AI model can include you as a source without introducing inaccuracy into its response.

• Use Organisation and Product schema consistently across your site

• Ensure your "About" and leadership pages are factually precise and regularly updated

• Build citations from credible external sources in your sector, not just link-building sources

• Align how you describe your category across all owned channels

What to Prioritise Before the Next Issue

Three Practical Actions This Week

Given everything above, here is what we would focus on if we were in your position right now:

1. Run a citation audit across ChatGPT, Perplexity, and Google AI Overviews for your five most important topic areas. Understand where you are being cited, where you are absent, and who is appearing instead of you.

2. Identify your two or three strongest existing content assets on topics where you have real expertise, and improve them with updated data, clearer structure, and explicit answers to the questions AI models are likely to surface.

3. Check your structured data implementation using Google's Rich Results Test and a manual schema validator. Missing or broken markup is still one of the fastest wins for AI visibility improvement.

What We Are Watching Into Q3 2026

A few forward-looking signals worth flagging. OpenAI has indicated further expansion of ChatGPT's real-time web access across more regions and query types through mid-2026. If that rollout continues as outlined, the volume of queries where ChatGPT actively selects sources will increase significantly, raising the stakes for brands not yet tracking their AI visibility.

We are also watching the emergence of AI-native search interfaces in markets outside North America and Western Europe. Adoption curves in Southeast Asia, Latin America, and the Middle East are accelerating faster than most Western-focused reports acknowledge. Brands with international footprints should be running visibility programmes that account for model behaviour across languages and regions, not just English-language queries.

The businesses that build their AI visibility now, before their categories consolidate around a small pool of trusted sources, are the ones that will be hardest to displace later. That is not hype. That is just how source selection in AI models works: once a model associates a brand with authority on a topic, the reinforcement loops are strong.

If you want to see how your site is currently performing across the 13 optimisation layers we track at Lua, start your assessment at luarank.com. The next issue of the Cited newsletter will be in your inbox on 9 May 2026.

Sources: BrightEdge AI Search Report Q1 2026; Semrush State of AI Search 2026; Google Search Central documentation updates, April 2026; Perplexity AI product changelog, Q1 2026; OpenAI product roadmap announcements, March 2026.