Cited Newsletter Issue 24 (May 19th 2026)

Welcome back to Cited, our weekly briefing on AI search visibility. Issue 24 covers a busy fortnight in the world of AI models and how they surface brand content. If you've been watching your referral traffic from ChatGPT or Perplexity flatten out, read on. There are some structural shifts this week worth paying attention to.

What's Changing in AI Search This Week

Google AI Overviews Are Expanding Source Diversity

We've been tracking this for three weeks now, and the pattern is clear. Google AI Overviews are pulling citations from a broader set of domains than they were in Q1 2026. The dominance of high-DA editorial sites is loosening slightly. Brand-owned content that follows strong structured data practices and clear topical authority signals is appearing more often in the source panels, particularly for mid-funnel informational queries.

The implication for marketing teams is straightforward: if you've been holding off on FAQ schema, HowTo markup, or entity-based content structuring because "that's an SEO thing," reconsider. These signals are feeding directly into AI Overview source selection. According to research from Search Engine Land, structured data implementation correlates with a 23% higher likelihood of being surfaced in AI-generated answers compared to unstructured equivalents.

Perplexity's Citation Behaviour Has Shifted Again

Perplexity updated its Pro product in early May 2026 with changes to how it prioritises sources for answer synthesis. The practical effect is that recency now weighs more heavily against domain authority for topics with a short shelf life. If you publish content in a fast-moving category (B2B SaaS, fintech, healthcare tech) and that content isn't dated, structured, and updated consistently, you're at a disadvantage against fresher sources from smaller sites.

One counterargument worth acknowledging: some marketers push back on the idea of constant content refreshing, arguing it dilutes editorial quality and creates maintenance overhead. There's merit to that concern. Our view is that selective refreshing (targeting the specific pages that are already in AI consideration sets but losing recency points) is far more efficient than a blanket "update everything" approach. The data from Lua's 13-layer website assessment consistently shows that 15-20% of a brand's content drives the vast majority of its AI citation potential. That's where refresh effort belongs.

ChatGPT Browse Is Now Treating Author Signals Differently

OpenAI's browsing capability appears to be placing more weight on named authorship and credentials when evaluating the trustworthiness of source content. Pages with identifiable authors, author bio pages, and linked professional profiles (LinkedIn, institutional profiles, published works) are seeing stronger citation rates in our tracked brand set. This aligns with what Moz reported on E-E-A-T signal evolution earlier this year.

If your content is still published under a generic "Admin" or "Team" byline, fix that. It's a two-hour task that has measurable impact on AI citation rates within 30 days based on the brands we're tracking.

Data Snapshot: AI Visibility Performance Across Lua's Brand Set

Every issue we share aggregate anonymised data from the brands running on Lua's platform. Here's the May 19th snapshot.

The gap between average and top-quartile brands consistently comes down to execution cadence. Brands in the top quartile are completing 85% or more of their scheduled Lua tasks each week. Brands at the average are completing around 50%. Consistency beats brilliance in AI visibility work. The models reward sustained signals, not one-off publishing bursts.

What the Top-Quartile Brands Have in Common

• They publish at least two pieces of entity-rich, structured content per week (not volume for its own sake, but content that explicitly builds brand and topic entity associations)

• They've completed their schema implementation tasks in full, including FAQ, Organization, and BreadcrumbList markup

• They monitor competitor citation sets monthly and adjust their topic coverage to plug gaps

• Their author profiles are live, linked, and cross-referenced across owned platforms

What We're Watching Into Q3 2026

Multi-Model Visibility Is Becoming Non-Negotiable

Six months ago, most marketing teams were focused almost entirely on ChatGPT visibility. That's understandable given its market share. But the picture is shifting. Perplexity's enterprise traction is accelerating (the company reported a 300% year-on-year increase in enterprise API usage in Q1 2026 according to TechCrunch), and Claude's adoption inside corporate workflows means your buyers may be getting AI-assisted answers from a model you've never optimised for.

Tracking visibility across a single model and calling it "AI search performance" is increasingly a blind spot. We track across ChatGPT, Perplexity, Google AI Overviews, and Claude by default for exactly this reason.

The Competitive Land Grab Window Is Narrowing

We've said this before, but the data keeps reinforcing it. Categories where early movers have built consistent AI citation programmes are starting to show compounding returns. Brands entering now, in May 2026, still have a meaningful advantage over brands that wait until Q4. But that window isn't infinite. Semrush's 2026 AI Visibility Report found that category leaders in AI citations are consolidating their positions at roughly twice the rate of traditional organic search, precisely because AI models tend to reinforce existing citation patterns once they're established.

If your team is still in "evaluating whether this matters" mode, the answer is yes, it matters, and the cost of waiting is measurable. The brands we onboarded in January 2026 are already building citation moats that will be difficult for latecomers to overcome in their categories.

Practical Task for This Week

Pull up your five highest-traffic blog posts. Check whether each one has: a named author with a live bio page, at least one FAQ schema block, and an internal link to your core product or service pages. If any of those three things are missing, fix them before you publish anything new. You're optimising existing AI consideration-set candidates, not starting from scratch.

That's Issue 24. We publish the Cited newsletter every Monday. If a colleague forwarded this to you, you can subscribe at luarank.com. See you next week.