Cited — 2026-04-27

Welcome to this week's edition of the Cited newsletter. Every issue, we pull together the most important developments in AI search, distil what they actually mean for marketing teams running real programmes, and cut through anything that doesn't move the needle. This week: model citation behaviour is shifting again, Google's AI Overviews are surfacing new source patterns, and there's a strategic window opening up that most brands are still missing.

What's Changed in AI Citation Behaviour This Week

ChatGPT Is Getting More Selective About Sources

Data from our platform across 40+ brands shows a clear pattern emerging over the past two weeks. ChatGPT is becoming measurably more selective about which sources it cites in response to commercial and informational queries. Specifically, we're seeing a drop in citations for pages that lack what we'd loosely call "citation anchors": clear author credentials, structured data markup, and first-person evidence (case studies, original data, named examples).

Pages that were ranking comfortably in ChatGPT responses three weeks ago have slipped, while competitors with stronger E-E-A-T signals have moved in. This isn't speculation. We track this across models every week, and the movement is consistent enough to call a trend rather than noise.

What to do right now:

• Audit your top-performing pages for named authorship and author bio markup

• Add structured data (specifically Article and Person schema) where it's missing

• Introduce original data or first-hand examples on pages targeting high-value queries

Perplexity's Source Rotation Is Accelerating

Perplexity's citation pool is rotating faster than it was six months ago. Sources that appear in one week's responses aren't guaranteed to persist into the next, particularly in competitive verticals like SaaS, finance, and professional services. The brands holding citation position consistently are those publishing on a structured cadence (at minimum, two to three substantive pieces per month) and actively building what we track as "citation authority" across third-party publications.

If you're only optimising your own site and ignoring external placement, you're leaving a significant visibility gap open for competitors to fill.

Google AI Overviews: New Source Patterns Emerging

Google's AI Overviews (the feature formerly known as SGE) are now pulling from a noticeably broader source pool than they were at launch. Early data suggested Overviews heavily favoured established publishers and high-domain-authority sites. That's still partially true, but we're now seeing mid-market brands appear in Overviews for specific long-tail queries, particularly where those brands have well-structured comparison content or detailed how-to pages.

The implication: if you've been assuming AI Overviews are out of reach because you're not a major publisher, that assumption needs revisiting. The window is open. It won't stay that way indefinitely as more brands catch on.

The Strategic Opportunity Most Brands Are Still Missing

Here's the honest picture. The majority of mid-market businesses are still treating AI search as a "watch brief" rather than an active programme. They're monitoring, reading newsletters like this one, and waiting for the channel to mature before committing resources. That's understandable. It's also exactly how you end up playing catch-up in twelve months.

The brands we work with that have achieved first-page ChatGPT rankings in under 40 days didn't do anything exotic. They followed a structured programme, executed consistently, and got there before their competitors started paying attention. That window exists right now in most verticals. It won't exist indefinitely.

A Counterargument Worth Taking Seriously

Not everyone agrees that AI search warrants prioritisation over traditional SEO right now. The counterargument, which we think is worth engaging honestly: AI search still drives a fraction of the referral traffic that Google organic delivers, and investing heavily in an emerging channel carries real opportunity cost. If your traditional search programme is underperforming, fixing that first is a legitimate strategic call.

Our view: these aren't mutually exclusive. A well-executed AI visibility programme reinforces your traditional SEO because the underlying signals (structured content, authority, expertise signals) overlap substantially. You're not choosing between channels. But if you're stretched thin and have to prioritise, fix the fundamentals first and layer in AI search systematically rather than reactively.

What the Next 12 Months Look Like

We'll go out on a limb here. By early 2027, AI search citations will function similarly to backlinks in 2012: the brands that built early, structured authority in these models will hold compounding advantages that are genuinely difficult for late entrants to close. The brands that waited will face the same dilemma SEO latecomers faced a decade ago: paying more, working harder, and getting less in return.

We're already seeing early signs of this compounding effect in our platform data. Brands that have been running structured AI visibility programmes for three to four months are maintaining citation positions with less active effort than they needed in month one. The model has learnt to trust them.

This Week's Recommended Actions

Immediate (This Week)

• Run a citation check on your three highest-priority query clusters across ChatGPT and Perplexity. Are you appearing? Who is? What do their pages have that yours don't?

• Check whether your key landing pages have author markup and Article schema in place. If not, this is a quick win with measurable impact.

• Identify one piece of content you can upgrade with original data, a named case study, or a first-person example. Schedule it for this week, not next quarter.

Medium-Term (Next 30 Days)

• Build or audit your external citation footprint. Are credible third-party sites referencing your brand in the context of your core topics? If not, a structured outreach and contribution programme is worth prioritising.

• Map your content against your top 20 AI search queries. Identify gaps where no page on your site adequately addresses what the model needs to cite you.

• Set up a basic tracking framework to monitor citation position across models weekly, not monthly. The pace of change requires weekly visibility at minimum.

Looking Further Ahead

The brands that will own AI search in 2027 are building their programmes in 2026. That means structured execution, not ad hoc content bursts. It means tracking competitor citation positions, not just your own. And it means treating AI visibility as a channel with its own logic and requirements, not just an extension of what you're already doing in SEO.

The Cited newsletter exists to give you the intelligence to make those calls well. Every issue, we report what we're actually seeing in platform data from live programmes, not repackaged industry announcements. If something is shifting, you'll hear it here first, with enough context to act on it rather than just file it away.

Next issue drops 4 May 2026. In the meantime, if you want to see where your brand currently stands across ChatGPT, Perplexity, Google AI Overviews, and Claude, Lua's 13-layer assessment will tell you exactly what's working, what isn't, and what to fix first. Most teams have their initial report within 24 hours.

Sources referenced:

• Google Search Central Blog, AI Overviews update documentation (April 2026)

• Search Engine Land, "How ChatGPT selects sources for citation" (March 2026)

• Perplexity AI product changelog and source attribution notes (Q1 2026)

• Lua platform data, aggregated citation tracking across 40+ brands (April 2026)