Cited Newsletter Issue 26 (May 21st 2026)

Welcome back to the Cited newsletter. Issue 26 covers the signal changes, platform shifts, and tactical updates that matter most to marketing teams building AI search visibility right now. We keep this tight and useful. No padding, no hype.

This week: a meaningful shift in how ChatGPT handles brand citations, why Perplexity's source diversity is narrowing, and what we're seeing in our own platform data across the 40+ brands currently running Lua programmes.

What Changed This Week in AI Search

ChatGPT Citation Behaviour: A Detectable Shift

Over the past two weeks, we've tracked a noticeable change in how ChatGPT surfaces brand references in commercial queries. The pattern is consistent across multiple verticals: structured FAQ content and clearly attributed expert statements are getting pulled into responses at a higher rate than generic product copy.

This isn't a confirmed algorithm update from OpenAI. It's a behavioural pattern we're detecting through visibility tracking. But the direction is clear enough to act on. Brands with well-structured, question-answer formatted content on their core pages are appearing in more ChatGPT responses for mid-funnel queries. Brands with dense, paragraph-heavy service pages are losing ground to those with cleaner, more extractable content structures.

The practical implication: if you haven't audited your key landing pages for AI extractability, that work belongs in your next sprint. This is one of the 13 optimisation layers we assess in every Lua scan, and right now it's one of the highest-impact areas to address.

Perplexity Source Diversity Is Narrowing

Perplexity has historically drawn from a wider spread of sources than Google's AI Overviews. That appears to be changing. Based on query sampling across sectors including SaaS, professional services, and e-commerce, the platform is increasingly favouring sources with strong domain authority, consistent publication frequency, and clear authorship signals.

The counterargument here is worth acknowledging: some marketers argue that Perplexity's narrowing is temporary, tied to a recent index refresh rather than a structural change. That's plausible. But even if it reverses, the underlying lesson holds. Publishing authoritative, author-attributed content with consistent cadence is a durable strategy regardless of which model you're optimising for.

For brands newer to AI visibility, this is a useful reminder that GEO (Generative Engine Optimisation) is not a single-platform game. Visibility across ChatGPT, Perplexity, Google AI Overviews, and Claude requires different content signals, and treating them as interchangeable is one of the most common mistakes we see.

Google AI Overviews: What's Working in May 2026

Google's AI Overviews continue to evolve, and the data from our tracked brands shows a few clear patterns for this month:

• Structured data markup (particularly FAQ and HowTo schema) is strongly correlated with AI Overview inclusion for informational queries.

• Brands with recent content updates (within the last 60 days) are outperforming older, static pages even when the older pages have stronger backlink profiles.

• Internal linking to a clear topical cluster is influencing which pages get pulled into Overviews for broader head terms.

One thing worth watching: Google is increasingly attributing overviews to sources that have demonstrated E-E-A-T signals at the author level, not just the domain level. That means author bio pages, LinkedIn profile links, and publication history are becoming ranking factors in a very practical sense. If your content is published as generic "team" bylines, that's worth revisiting.

Lua Platform Data: What We're Seeing Across 40+ Brands

Median Time to First-Page ChatGPT Visibility

We track visibility milestones across every brand on the platform. Here's a snapshot of performance data from our active cohort as of May 2026:

The correlation between task completion rate and visibility outcomes is not subtle. Brands in the top quartile are completing more of their scheduled tasks each week. That's it. The programme works when it's followed. The biggest gap we see isn't strategy, it's execution consistency.

The Brands Winning Right Now

The brands seeing the fastest AI search visibility gains share a few common traits. They've assigned a dedicated owner (typically a marketing director or growth lead). They treat their Lua execution calendar as a non-negotiable weekly commitment, not an optional sprint. And they review their competitor benchmarking data fortnightly rather than monthly.

That last point matters more than it might seem. AI search visibility shifts quickly. A competitor who wasn't ranking in ChatGPT for your core queries in March 2026 might be appearing consistently by May. Brands that catch these shifts early can adjust their content approach before the gap becomes a structural disadvantage.

What to Prioritise This Week

For Teams Mid-Programme

If you're 4 to 8 weeks into your Lua programme, the highest-leverage work right now is schema completion and internal link architecture. Both are showing outsized impact on Google AI Overviews inclusion this month. Check your task calendar. If these haven't been scheduled yet, flag them to move earlier in your queue.

For Teams Just Getting Started

The 13-layer scan will surface your biggest gaps. But before you run it, do a quick manual check: can you find your brand name in a ChatGPT response to your three most important customer queries? If you can't, that baseline tells you everything about where you're starting from.

A Note on the Competitive Window

We're often asked whether the AI search opportunity is still early enough to be worth pursuing. The honest answer is: yes, but not indefinitely. Across most sectors we track, fewer than 15% of mid-market businesses have an active AI visibility programme in place. That window won't stay open at the same width for long. The brands building structured programmes now are establishing citation patterns and source authority that will be significantly harder to displace in 12 to 18 months.

The future direction is also worth naming clearly. As AI models move toward agentic behaviour (handling bookings, purchases, and decisions autonomously), the brands already embedded as trusted sources in model memory will have a structural advantage that goes well beyond search rankings. Getting cited today is not just about traffic. It's about positioning for a version of commerce where the AI is the buyer's first and sometimes only touchpoint.

That's Issue 26. Next edition lands May 28th. If you have data, observations, or questions you'd like us to address in the Cited newsletter, reply directly. We read everything.

Sources:

• OpenAI usage and response behaviour documentation, openai.com

• Google Search Central, AI Overviews guidance, developers.google.com

• Perplexity AI source attribution patterns, perplexity.ai

• Lua platform visibility tracking data, internal cohort analysis, luarank.com