For the past two decades, appearing in search results meant one thing: ranking on Google. The mechanics were well understood — build authoritative content, acquire inbound links, optimise for the right keywords, earn a position on the first page. The system was competitive and imperfect, but its logic was consistent.
That logic is shifting, and the shift is happening faster than most organisations have noticed.
AI-referred sessions — visits that arrive via ChatGPT, Perplexity, Google’s AI Overviews, and similar platforms — jumped 527% year-over-year in the first five months of 2025. Google’s AI Overviews now appear on roughly 16% of all searches, up from just 6.49% at the start of the year. ChatGPT, as of April 2026, has more than 900 million weekly users. These are not marginal numbers. They describe a fundamental change in how a significant and growing portion of the population looks for information, makes decisions, and discovers organisations like yours.
The difference between traditional search and AI search is not merely technical. In traditional search, a user types a query and receives a ranked list of links. They click. They arrive on your site. The visit is recorded. In AI search, a user asks a question and receives a synthesised answer drawn from multiple sources — often without clicking through to any of them. AI answers drive measurable increases in branded queries and assisted conversions, even when raw session counts stay flat. Your content may be influencing decisions you can’t see in your analytics.
This creates a new question for any organisation that relies on being found online. It is no longer sufficient to ask: does my content rank? The question is: does my content get cited?
The distinction matters because the signals that drive citation are somewhat different from the signals that drive ranking. Traditional SEO rewards keyword optimisation, backlink volume, and page authority. AI systems favour content that is clear, specific, well-structured, and demonstrably authoritative — content that a language model can extract a coherent answer from without ambiguity. Research has found a systematic bias in how AI systems source information, favouring earned media — third-party, authoritative sources — over brand-owned content. An organisation cited in a credible industry publication is more likely to appear in an AI-generated answer than one whose content exists only on its own site.
What this means practically is that the fundamentals of good content — specific, accurate, well-written, clearly structured — matter more than ever, not less. Content that exists primarily to rank, padded to hit a word count and stuffed with keywords, is exactly what AI systems are designed to filter out. The content that surfaces in AI answers is content that actually answers a question well.
For most organisations, the immediate priority is not to abandon SEO but to understand that it is now one part of a wider picture in the content strategy. The same content that ranks in traditional search can, if it is genuinely useful and properly structured, be cited in AI-generated answers. The two objectives are compatible. Schema markup, clear information hierarchy, and specific factual claims all help both.
The more significant shift is cultural rather than technical. Organisations that have treated their content as a ranking exercise — producing material designed for an algorithm rather than for a reader — are going to find AI search less forgiving than Google has been. The systems making these decisions have been trained on vast quantities of human writing, and they are reasonably good at distinguishing between content that knows what it’s talking about and content that is performing knowledge.
The organisations that will navigate this transition well are the ones producing content with genuine authority in their field. That has always been the correct approach. It is now, increasingly, the only one that works.
