SEO in the GEO Era

Search did not die in 2024 when AI Overviews launched, and it did not die in 2025 when ChatGPT Search opened up. But it did fundamentally split. Every UK brand now competes for two distinct surfaces at the same time: the classic ranked-links result, which still exists and still drives revenue, and the cited-answer result, which sits above it and is increasingly where the discovery actually happens.

The teams that have grasped this are quietly reorganising their entire organic discovery operation around it. The teams that have not are writing more blog posts.

The state of UK organic discovery in 2026

Three structural facts now define the discipline:

• Informational click-through is collapsing. When the SERP returns an AI Overview that answers the user's question, the click-through to organic results below it falls dramatically. UK publishers reporting 30–50% organic traffic losses on informational content are not exaggerating. They are describing a structural reallocation, not a temporary dip.
• Commercial-intent traffic is largely intact. Users still click when they are about to buy. SERPs for "best AI marketing agency UK" or "Klaviyo migration cost" still send qualified traffic to the top results because users want to compare options, not absorb a synthesised summary.
• A new discovery channel exists. ChatGPT, Perplexity, Claude, and Gemini are now genuine, measurable referral sources for UK brands. Their referrer share is small relative to Google, but their consideration-stage influence is much larger than the click share suggests, because they are being used to shortlist before the user ever Googles a brand name.

What's actually changing

The shift inside organic discovery comes down to four practical changes in how the discipline is operated:

1. From keywords to entities. Optimising for keyword strings is being replaced by optimising for entity recognition: does the LLM know what your brand is, what category you are in, what you are credible on? Entity clarity is now the single biggest predictor of citation likelihood.
2. From content volume to content density. Long, fluffy, keyword-stuffed posts perform worse — both with crawlers and with LLMs — than shorter, denser, fact-rich pieces. The "10x content" playbook from 2018 is now actively counterproductive.
3. From anonymous to named. Schema.org Person markup with genuine expertise signals (LinkedIn, prior published work, affiliations) is now a measurable input into citation likelihood. Anonymous content gets cited far less than equivalently-good named content. This is structural, and it favours founders and senior operators who can credibly attach their name.
4. From link-building to citation-building. The metric that matters is no longer "how many backlinks did this earn" — it is "how often is this content quoted by LLMs and on the sources LLMs train on". Those are overlapping but distinct things. The latter requires PR-style placement work, not link-building tactics.

Live UK signals

• UK publishers with strong original data programmes (research reports, proprietary indices, surveys) are seeing their citation rate in ChatGPT and Perplexity grow month over month, even as their Google organic traffic shrinks. The underlying revenue is stable or growing.
• Pure-play SEO agencies in the UK are reporting margin pressure as clients question the value of traditional content and link-building work. The agencies adapting are repositioning as "organic discovery" practices that include GEO, original research, and PR-style citation building.
• Schema.org adoption — particularly Person, Organization, and Article markup — is now a measurable factor in AI Overview inclusion. Sites with comprehensive schema appear in AI Overview citations meaningfully more often than equivalent sites without.
• Reddit, Wikipedia, and a small number of established UK publications are now disproportionately cited by major LLMs. Brand presence on these surfaces — earned, not paid — has become a leading indicator of future citation share.

What this means for UK CMOs

1. Audit the click-share collapse honestly. Pull your last 12 months of organic traffic by query type. The informational queries are probably down. The commercial queries are probably steady. That tells you which content categories to retire and which to defend.
2. Reorganise the team around organic discovery, not SEO. SEO as an isolated function is too narrow for the work that now matters. The team needs writers, researchers, PR-minded operators, and technical SEO depth — under one strategy.
3. Invest in original data. The single highest-leverage move in 2026 is publishing proprietary research and benchmarks. It compounds for years, gets cited by LLMs, earns backlinks organically, and gives the PR team something real to work with.
4. Put names on everything. Anonymous content is increasingly invisible to LLM citation. Senior leaders should be bylined, with proper schema and genuine expertise pages.

For the broader strategic context, see The End of the Marketing Funnel and Who Owns the Customer in an AI World?.