Over the past two years the path by which people arrive at web pages has shifted sharply. Pew Research Center tracked browsing behavior for 900 U.S. adults and found that when Google shows an AI summary, users click a traditional search result only 8% of the time — roughly half the 15% click rate when no summary appears. Links cited inside AI answers perform worse still: users click those at about a 1% rate.
Publishers have felt the effects. Chartbeat data reported by Axios shows page views from Google Search fell 34% across its publisher network between December 2024 and December 2025; small publishers lost roughly 60% of search referral traffic over two years. Business Insider’s organic search traffic dropped 55% over three years, and some smaller outlets have already closed. Chatbot referrals have grown (more than 200% year-on-year) but still account for less than 1% of publisher page views.
Machines read more even as humans click less
Similarweb’s 2026 Generative AI Landscape report highlights that while AI platforms send fewer humans to web pages relative to the answers they create, the AI systems themselves are consuming the web at an increasing rate by fetching live web pages to answer user questions. The share of ChatGPT answers that included live web citations grew more than fivefold in under a year, reaching 6.8% of all answers by May 2026; in some categories, such as travel, that figure is as high as 22.6%.
All major AI search products fetch live pages from indexes and synthesize answers from them, so the quality of AI responses depends on the health of the underlying content layer. As Lily Ray, VP of SEO and AI search at Amsive, notes in the Similarweb report, if your organic visibility falls, your AI search visibility tends to follow because models are less likely to find your content.
This dynamic creates a troubling feedback loop: AI answers are built on an information supply chain whose funding model — ad-supported clicks — is eroding largely because AI answers are replacing those clicks. A critical question for the future of the open web is whether viable alternative business models will emerge, and what they will look like.
A replacement economy is forming inside chat
There are early signs of how the market may shift. After ChatGPT’s May 7, 2026 search update, which surfaced prominent clickable brand links inside answers, referral traffic from ChatGPT jumped 157% in a week. The shape of that traffic changed: the share of referrals landing on homepages more than doubled, from about 25% to nearly 60%.
Traditional search used to send users to specific articles and deep pages tied to queries. AI referrals increasingly deliver a pre-informed visitor to a brand’s front door: the chatbot did the research and comparison, and the person arrives ready to act. Similarweb’s data shows AI-recommended brands receive two to four times as many subsequent visits as competitors that were not recommended.
Money follows behavior. Similarweb’s ad-intelligence data shows sponsored results appeared in 26% of U.S. desktop ChatGPT conversations in June 2026, up from 14% a month earlier. Two-thirds of those ads appear after the second prompt and are targeted on conversation context rather than a keyword. Click-through rates are around 0.50%.
The traditional search-engine keyword auction is being supplanted by paid placement inside conversations, targeted on accumulated context — a direct challenge to the auction model Google has run and dominated for two decades.
Google is both incumbent and disruptor
Google is not merely a bystander: it is simultaneously the incumbent being disrupted and one of the largest players driving disruption. AI Overviews now appear in a growing share of Google searches — according to Similarweb, more than 40% by May 2026 — and visits to Google’s conversational AI Mode have steadily climbed. In effect, Google is cannibalizing parts of its own click economy.
Yet the ground under Google’s core business is shifting. eMarketer projects Google’s share of U.S. search advertising will fall below 50% in 2026 — the first time since roughly 2004. Much of that lost share is going to Amazon, whose sponsored product searches count as search advertising and are growing three times faster than Google’s. Conversational ads are a small part of those totals today, but they open a second front in a market Google has not had to fight on before.
Structural legal changes are also in motion: a federal court entered final judgment in the Department of Justice search antitrust case in December 2025, imposing remedies that bar exclusive default agreements and require Google to share search data with qualified competitors. Google appealed in January 2026; the DOJ cross-appealed seeking stronger remedies. Regardless of the appeals’ outcome, the de facto arrangement that made Google the web’s tollbooth — defaults everywhere and a closed index — is ending just as conversational advertising reshapes the landscape.
The competitive field that emerges is genuinely new. OpenAI, Google, Perplexity and Microsoft are now competing not only for users but also for the advertising demand that funded the open web, and none of them — including Google — controls the new surface the way Google controlled the old one.
Does conversational advertising help or hurt the open web?
It’s not yet clear whether the new model benefits or harms the web. The web as a destination for human attention is shrinking, putting ad-supported publishers in real trouble. At the same time the web as a machine-readable substrate is growing in importance, and a referral and advertising economy is forming that routes value to brands rather than to content pages.
Publishers may have limited influence over how this plays out. Ahrefs, analyzing over a billion data points across its studies, found that 67% of ChatGPT’s most-cited sources are things marketers cannot easily influence; Wikipedia alone accounts for nearly 30%. Moreover, 28.3% of ChatGPT’s most-cited pages have zero Google organic visibility, suggesting the retrieval layer is only partially tethered to traditional search — a complication for anyone assuming SEO success will automatically translate into machine citation.
Rebuild your site for visitors who arrive from conversations
Three independent datasets show that in the new world the pages AI systems cite and the pages AI systems send humans to are different pages performing different roles. That is a significant change and many teams are still optimizing for the old environment.
Similarweb finds 65% of ChatGPT-cited URLs sit two or three folders deep in a site, while 58.8% of referral traffic lands on homepages. Ahrefs observed a similar split: more than 80% of its AI referral traffic goes to its homepage, product pages and free tools rather than editorial content. A Previsible analysis of 6.77 million AI-referred sessions found a third common destination: 28.8% of ChatGPT referrals land on internal site search pages — a navigation surface many publishers have long neglected because Google searches used to provide that function.
The practical action is to audit your search traffic patterns: pull AI referral logs and any citation data you can access, and map which pages are quoted as evidence versus where visitors actually enter. If your site is operating in the new environment, three clear steps follow:
- Structure deep pages, documentation, comparisons and benchmarks so they are citable: state specific claims, use clear headings and descriptive URLs. (Ahrefs found pages with natural-language URL slugs are cited at 89.78% versus 81.11% without.)
- Rebuild the homepage for a visitor who arrives with context from a conversation rather than from a blue link: these users already know you likely have what they need, so get them to it quickly.
- Invest in internal search: formerly neglected, it is now an acquisition surface that deserves real UX attention.
Much remains uncertain and in flux, but every independent source that has looked finds the same trend: the old click economy is not coming back, and those entities that learn to be quoted by machines and to convert the humans those machines send will control whatever the web becomes next.
