Search engines used to send readers to publisher websites. Today, they are increasingly designed to keep them. Referrals from Google Search to hundreds of publishers fell 40 percent from July 2025 to July 2026—double the rate of decline seen the previous year, according to Chartbeat data reported by Gretel Kahn at the Reuters Institute. The industry’s greatest fear now has a name: Google Zero, the point at which search ceases to be a doorway to the broader web and becomes an end-to-end answer engine.
While search engines have traditionally refrained from placing AI Overviews directly on fast-moving news queries, news-adjacent topics are a different story. Background context, historical summaries, and politician biographies routinely trigger AI-generated answers that directly replace the explainer pieces newsrooms relied on for search traffic. A Pew study analyzing more than 68,000 searches by 900 U.S. users revealed that only 1 percent of searches featuring an AI Overview resulted in a click to a cited source. Being cited in a response is no longer synonymous with being visited.
Industry leaders are being clear-eyed about the changing landscape. Ezra Eeman, director of AI in media at WAN-IFRA, observes that citation preferences fluctuate unpredictably between sources like Wikipedia, Reddit, and YouTube across different platforms. Shelby Blackley at The Athletic views the underlying algorithms as a black box that changes too frequently to game through traditional tricks. Audience consultant Clara Soteras argues that publishers must update their success metrics altogether: because a mention in an AI overview does not generate a page view, measuring success purely by unique visitors means tracking a bygone distribution model.
To adapt, media organizations are shifting their strategies. Ben Berkowitz at Axios is focusing on high-authority reporting formatted cleanly for machine parsing, alongside direct-to-reader channels like newsletters and live events that build long-term retention. Jessie Willms at The Guardian treats generative search optimization as an extension of traditional SEO rather than a total replacement, warning newsrooms against reshaping their editorial voice around volatile algorithmic trends.
At the same time, platforms are experimenting with direct licensing deals. As reported by Nieman Lab, Google’s News AI pilot expanded to encompass more than 400 publications, including national outlets like CNN, the Financial Times, Axios, and Forbes, as well as local papers like The Post and Courier and The Texas Tribune. Under this program, publishers receive compensation to license content for experimental AI features—such as automated story previews and audio briefings—rather than earning revenue from referral traffic. Crucially, these previews come marked with disclaimers about machine-generated content and potential errors. These arrangements represent content licensing, not the organic referral engine around which digital publishing was originally built.
The journalism capable of thriving in this environment is the journalism an AI summary cannot replicate: original local reporting, real-time updates, and trusted brand identity that brings readers back by name. Showing up where questions are asked is important, but newsrooms must not mistake a search citation for a real reader visit.
Our mission is to help newsrooms navigate this shift by providing tools that strengthen direct audience relationships. By enabling publishers to build rich, interactive experiences on their own platforms, we help media companies engage readers directly on their own terms—ensuring their journalism is read in full, rather than summarized away.