We’ve been here before. When Google launched Panda, SEO was dead. When voice search arrived, SEO was dead. When Featured Snippets took position zero, SEO was dead. Now AI Overviews are reducing click-through rates on position-one results, and the same article is being written again with new stats in it.
The discipline is not dying. The surfaces are changing. Those are different statements with completely different implications for what you should do next.
What actually stays constant
The core question of SEO has been the same since 1998: where does your audience go to find information, and are you the answer they find? That question doesn’t change when the surface changes from ten blue links to voice responses to AI-generated answers to agentic recommendations. The execution adapts. The goal is identical.
The 3Cs framework I developed in 2010 — Code, Content, Contextual Linking — still describes the technical foundation of every site that ranks, regardless of platform. The Dog Walker Portsmouth site I built and optimised in 2009 has held its number one position for over seventeen years through Panda, Penguin, Hummingbird, RankBrain, BERT, Core Updates, and now AI Overviews. Not because I was chasing each algorithm change. Because the fundamentals were right and have stayed right.
The Hair Lounge Totton site survived two domain migrations and a full rebrand. Eco Montessori ranks number one nationally. Azure Outdoor Living scaled to seven-figure turnover. None of those outcomes came from pivoting to the latest SEO trend. They came from doing the foundational work properly and building on it.
What the doom crowd keeps misreading
The statistic doing the rounds right now is that click-through rates at position one drop by around 58% when an AI Overview is present. That’s real data from Ahrefs. What the doom articles don’t mention is that the same Seer Interactive research tracking 12 million website visits found that traffic arriving via AI citation converts at 14.2%, compared to 2.8% for traditional organic traffic. AI-referred visitors are five times more likely to become customers.
The pie isn’t shrinking. The route to the pie is changing. Businesses that only optimised for clicks from ten blue links are being displaced. Businesses that are cited by AI systems as specific recommended providers are capturing higher-intent traffic with dramatically better conversion rates. That is not a death. That is a redistribution — and the redistribution is happening right now, while the people writing “SEO is dead” are still writing about it instead of building the infrastructure that determines who gets cited.
The new surface requires a new layer, not a new discipline
Here’s what has changed: the discovery surfaces now include AI Overviews, Perplexity, ChatGPT Search, Copilot, and agentic systems making recommendations without a human query at all. Getting your content retrieved and cited on these surfaces requires structured content, entity corroboration, and CITATE-compliant page architecture — work that traditional SEO practice didn’t explicitly address.
But the foundation is identical. You need Bing indexing for ChatGPT Search and Copilot — and Bing indexing is technical SEO. You need entity recognition for AI recommendation eligibility — and entity SEO is a natural extension of the on-page optimisation and link-building work you’ve been doing for years. You need CITATE-compliant content structure for extraction — and that’s just well-written, well-structured content with named sources and explicit claims.
The SEO practitioners who built strong technical foundations, genuine topical authority, and real off-site corroboration are the ones appearing in AI answers now. The ones who chased shortcuts, bought links, and produced thin content for keyword volume are the ones losing ground. That pattern is not new. It’s the same pattern as every previous surface change, just playing out on a new set of platforms.
What to do instead of panicking
Run the AI Discovery Stack diagnostic on your own site. Where is the leak? Is Bing indexing your key pages? Is your entity recognised and corroborated? Are your content sections structured for extraction? Is your brand established enough for AI systems to name you specifically? Those four questions produce an action list. The action list is SEO, applied to 2026 surfaces.
The discipline has an unbroken thread from directory submissions in 1998 to AI agent recommendations in 2026. Every practitioner who understood that thread — that the goal is always to be the answer your audience finds — has adapted without rebuilding from scratch. Every one who treated each surface as a new discipline has rebuilt six times and is about to do it again.
Don’t rebuild. Extend. The full argument is at The Unbroken Thread. The five-layer framework showing exactly where to extend is at the AI Discovery Stack.
