Tech · 22 August 2026 · 4 min read
Google Launches Preferred Sources Button to Combat AI-Driven Traffic Loss
In brief: In an effort to mitigate the decline in web publisher referral traffic caused by AI Overviews and synthetic summaries, Google has launched an embeddable «Preferred Sources» button for online publishers. The tool allows readers to bookmark trusted outlets directly from web pages to prioritize them across Google Search, AI Mode, and Discover, while feed personalization gains conversational natural-language prompt controls.
by Team Mocchi's
The rise of synthetic answers in web search has turned digital distribution into a high-stakes standoff between technology platforms and online publishers. With the rapid expansion of features like AI Overviews and conversational query modes, user click-through rates to standard web links have seen substantial drops across multiple verticals. In response to mounting friction with the media sector, Google has announced a suite of updates intended to restore referral pathways and give users more granular control over algorithmic recommendations.
At the center of this push is a new interactive component designed for online publishers to embed directly on their properties, connected to Google's ranking and content distribution engine.
An embeddable button to reinforce reader loyalty
As reported by TechCrunch, Google has introduced an embeddable «Preferred Sources» button that websites can place alongside their editorial content. When a reader clicks the button, the host publication is immediately added to their preferred sources list within Google Search, Discover, and Google News without navigating away from the page or opening account settings.
Designating a preferred source directly alters what appears in organic search snippets, Top Stories carousels, and AI Overviews. Google noted that users have already selected over 345,000 unique publications through earlier configuration menus, and internal metrics indicate that users are twice as likely to click through to a preferred source when it appears in search results.
Conversational prompt tuning lands on Google Discover
Alongside publisher-facing tools, Google is overhauling how readers manage Google Discover, the algorithmic content stream accessed by hundreds of millions of smartphone users daily. According to The Verge, the Google mobile app will soon allow users to steer their Discover feed by describing their interests in plain text using an interactive chatbot interface.
By accessing the settings menu on any card within Discover, users can input natural-language instructions—such as asking for more deep dives on hardware engineering while filtering out celebrity gossip. The system confirms the criteria and stores the preference for subsequent feed generations. Google is also rolling out similar personalization controls to daily audio briefings inside the Google News app on Android.
A pragmatic truce between AI search and the open web
These additions highlight the delicate balancing act defining modern search infrastructure. Google cannot halt the deployment of generative AI answers without ceding market share to competing conversational assistants, yet the quality of those answers relies entirely on an active, financially viable ecosystem of human creators and newsrooms.
Giving publishers a direct mechanism to turn occasional traffic into persistent algorithmic priority is an explicit compromise: rather than scaling back synthetic overviews, Google is offering established digital brands a protected lane within its AI-curated discovery channels.
Mocchi's take
The introduction of an embeddable preferred source toggle signals a fundamental shift for businesses and media teams running digital publications: traditional SEO targeting search keywords is no longer sufficient without explicit brand loyalty and direct user opt-ins. Implementing these interaction hooks early and delivering distinctive editorial value will be critical to remaining visible as search pages convert into AI-generated answers. From an architectural perspective, Google's prompt-driven Discover tuning also confirms that conversational natural-language interfaces are rapidly superseding static preference forms across modern recommendation engines.