Authored by : Lalit Sharma, Founder-CEO, PR Companion and Ethane Web Technologies
For about 20 years, marketers have treated ranking first on Google as the end goal of creating brand visibility. It was always simple: rank number one of Google, get the traffic on the website, it will generate leads & sales, and visibility will follow on its own. That equation has started to break down since generative AI entered search, and the scale of the shift is now visible in the numbers.
A study published in March 2026 tracked 863,000 keywords and 4 million AI Overview URLs. As per the study, merely 38% of pages cited in Google’s AI Overviews were ranked in the top 10 organic results for the same search term. While the number was at 76% just seven months ago. In less than a year, the link between ranking position and AI citation has roughly been cut in half. The question worth asking is why. What changed in that window was not how Google ranks pages. It was what AI systems are actually built to do with them.
Google’s ranking system and an AI system’s citation logic are answering two different questions. Google questions which page best matches the purpose behind a search, based on the relevance, backlinks and authority signals built up over years. On the other hand, an AI tool goes narrower. It questions which piece of content can be pulled out cleanly, trusted, and used to construct a direct answer. A page can be written to win on the first question and still fail the second. Long intros, storytelling build-up, and a key insight buried in paragraph twelve. All of this can still rank well on Google. None of it helps when an AI model is scanning for something it can lift and use immediately.
This is where the gap actually opens up. Earlier, the ranking algorithms were built on just one factor: a page as a whole, its backlinks, its age, its topical depth, and its overall authority on a subject. On the other hand, AI tools narrow it down, often a single paragraph or even a sentence, and ask if it can establish itself as a fact or an answer. A brand can still invest a huge amount in strong domain authority but consist of vague answers and rank on Google because of the SEO. However, will AI cite that page on its chat? It is the question that has come to the surface. Simultaneously, a lesser-known page will clear answers and facts, backed with sources and structural content, and has a true chance to appear in AI searches. In simple words, ranking rewards a page's overall weight. Citation rewards the clarity of a single passage within it.
In 2024, Gartner anticipated that traditional search engine volume would drop 25% by 2026 as generative AI tools held a growing share of queries that once were “Googled”.
Though the projection has not come to exactly the same, as Google still handles the overwhelming majority of search queries around the world. However, there has been an underlying shift. AI systems increasingly standing between a question and its answer are now well documented and measurable.
There is also a separate question worth asking: once a citation happens, what kind of content is actually being cited? In a Muck Rack report, ‘What is AI Reading’, the research suggests more than 25 million links cited by ChatGPT, Claude and Gemini. The finding has held steady across all three editions: earned media, meaning independent, non-paid coverage, accounts for 84 per cent of everything these AI systems cite. Paid or advertorial content barely registers, at 0.3 per cent. Journalism specifically makes up around 27 per cent of citations.
Put these three findings together and a pattern emerges. Ranking position is losing its grip on AI visibility. Traditional search volume itself is shrinking as a share of how people find answers. When AI cites something, it is the result of independent reporting done on the basis of third-party coverage far more often than owned brand content, no matter how well that content ranks.
It does not mean that search engine optimization is irrelevant. Factors like domain authority, technical health and well-structured content still matter and serve as a base for AI systems. The only change that has come is that ranking on Google is not enough on its own. Earlier, it was seen as the endgame, now it is where the foundation is built.
For anyone responsible for a brand's visibility, in marketing or in communications, the practical shift is in what gets measured and what gets prioritised. A page landing in position one is no longer proof that the job is done. Whether that brand is being cited by name in an AI-generated answer or where that citation is coming from – earned coverage, a forum thread, a comparison article, a product page – is becoming just as important to track. The data above suggests the businesses paying attention to that distinction now will be the ones still visible once the answer itself, not the link beneath it, becomes what most people actually see.
