Something has quietly broken in how people find information online, and most of us barely noticed it happening. You ask a question, a chatbot answers it in a tidy paragraph, and you move on with your day. No blue links, no scrolling, no visit to the website that actually did the work of researching and writing that answer. Multiply that moment by billions of searches a day and you get one of the biggest shifts in how the internet works since Google itself took over from directories like Yahoo in the early 2000s.
The Numbers Behind the Shift
Search behavior data from SparkToro shows that the share of Google searches ending in a click to any website has been falling for years, and the drop has accelerated sharply. When Google shows an AI Overview at the top of the results page, users click through to a traditional result only about 8 percent of the time, compared to roughly 15 percent when no AI Overview appears. Google’s newer AI Mode pushes that further still, with a large majority of sessions ending entirely inside Google’s own interface and never reaching an outside site at all.
Publishers are feeling this directly. Search traffic to publisher websites fell roughly a third globally over the past year, and even harder in the United States. Some sites built around straightforward “how to” content have reportedly lost the bulk of their organic search traffic in the span of a few quarters. Meanwhile, the chatbots themselves are growing at a pace traditional search never managed. ChatGPT traffic grew more than 600 percent year over year through mid-2025, and Perplexity grew well over 200 percent, while Google and Bing saw their own numbers slip slightly over the same stretch.
Why Gartner Thinks PR Budgets Are About to Double
This is where the story gets genuinely interesting for anyone who runs a business, a brand, or even a personal website. Gartner’s research team recently published a set of predictions for 2026 communications strategy, and the headline finding is that public relations and earned media budgets are on track to roughly double by 2027. That is not a modest bump. It is a fundamental reallocation of marketing spend, and it is happening because of how AI chatbots actually decide what to say.
When a chatbot answers a question, it is not running a paid ad auction the way Google Search does. It is pulling from a mix of sources it considers credible, and Gartner’s analysis found that more than 95 percent of the links cited by AI answer engines are unpaid mentions, with over a quarter of those tracing back directly to earned media coverage: news articles, expert quotes, reviews, and independent write-ups rather than sponsored placements. A company can no longer simply buy its way into the answer a chatbot gives. It has to earn a mention from a source the model already trusts.
That single fact is reshuffling marketing budgets across the industry. Reallocating money that used to go toward paid search and display ads into PR, media relations, and credible third-party coverage is quickly becoming the more rational play, since that is the exact kind of content large language models lean on when they answer a question.
Enter GEO, the Replacement for Old-School SEO
The industry has already coined a name for the discipline built around this shift: generative engine optimization, or GEO, sometimes also called answer engine optimization. Where traditional search engine optimization was about ranking a URL as high as possible in a list of results, GEO is about becoming the source an AI model actually cites or paraphrases when it builds an answer.
The two disciplines do not overlap nearly as much as marketers might hope. Research from one GEO analytics firm found that the overlap between the top links shown in a regular Google search and the sources an AI system chooses to cite for the same question has fallen from around 70 percent to under 20 percent, and that gap keeps widening. In practice, that means a page can rank perfectly well in classic search and still be invisible to the chatbot a growing share of your potential customers are now asking instead.
Instead of chasing keywords, teams doing this well now focus on things like clear, well-structured explanations that a model can lift cleanly, consistent facts about a brand or product across many independent sources, and genuine coverage from outlets and experts a model has learned to trust. It is less about gaming a ranking algorithm and more about being the kind of source a language model would reasonably quote if a person asked it the same question directly.
What This Means for Regular Internet Users, Not Just Marketers
If you are not running a business, this might sound like an industry problem that does not touch you directly, but it does in a couple of ways worth knowing about. First, the web you rely on for information depends on websites getting enough traffic to stay funded, whether through ads, subscriptions, or sales. When AI bots could soon outnumber actual human visitors online, and a chatbot answer keeps you from ever visiting the site that did the original research, that site earns less to keep producing that research in the first place. Fewer visits eventually means fewer of the independent reviews, guides, and reporting that both people and AI models depend on.
Second, it changes how much you should trust a quick chatbot answer at face value. These systems are summarizing and sometimes compressing information from sources you never see, and errors or outdated details can slip through in ways they would not if you were reading the original article yourself. If you have ever wanted to see the unfiltered search results rather than a synthesized summary, it is worth knowing you can still hide Google’s AI Overviews from your search results with a simple search trick, which at least puts the choice back in your hands.
If all of this AI search and chatbot terminology still feels a little fuzzy, our plain-English guide to what AI actually is is a good place to get grounded before diving deeper into how these tools are reshaping the web.
The Bigger Picture
None of this means traditional search is disappearing overnight, and it certainly does not mean marketing budgets are shrinking, quite the opposite. What it means is that the old playbook of ranking a page and buying a few ads is no longer enough on its own. The businesses adapting fastest are the ones treating credible, independently reported coverage as core infrastructure rather than a nice-to-have, because that is increasingly the only kind of content the next generation of search, one you talk to instead of type into, actually trusts.
