Google is rolling out three personalization features: natural language controls for Discover, an embeddable Preferred Sources button, and topic based audio briefings in Google News on Android. Discover users can describe what they want to see—or avoid—in ordinary language; Google may ask follow up questions, update...
Research answer

Create a landscape editorial hero image for this Studio Global article: What personalization features did Google announce for Search, Discover, and Google News—including the chatbot-style Discover controls that l. Article summary: Google announced three related personalization tools: conversational controls for Discover, an embeddable Preferred Sources button for publishers, and more configurable Android Google News audio briefings. They may deepe. Topic tags: general, general web, documentation, user generated, government. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text,
Google is expanding personalization across Search, Discover and Google News with three connected features: conversational controls for the Discover feed, an interactive Preferred Sources button that publishers can embed on their sites, and more configurable audio briefings in Google News for Android.
The strategy is clear: let people tell Google what they want, then make favored sources easier to find. For publishers, however, the promise is narrower. These tools may improve retention and engagement among readers who already have strong preferences, but the available evidence does not show that they can replace general Search referrals lost when AI-generated answers satisfy queries without a click.
In the Google app, users will be able to open a three-dot menu on a Discover item—or use a dedicated control—and describe the topics, sources or types of content they want to see more or less of. The interaction is designed to work like a conversation: Google may ask follow-up questions, then adjust the feed in response. The requests are also remembered so they can influence later recommendations.
Google says the feature is rolling out in the coming days, so availability may vary by user and region while the rollout progresses.
This changes the personalization model from simple reactions such as hiding a story or marking a topic as unwanted to an explicit instruction layer. Instead of waiting for the recommendation system to infer a preference, a user can state it directly in natural language.
Google is also introducing an interactive Preferred Sources button that publishers can place on their own pages. When a reader selects it, the publication is added as a preferred source in Google, and the reader is returned to the page they were viewing.
Google’s Search Central documentation recommends a JavaScript implementation for the interactive button. Publishers that cannot use that approach can use a deeplink to send readers to Google’s source-preferences interface.
A selected source can be surfaced more prominently for that user across Google experiences including Top Stories, AI Overviews and AI Mode. That makes the feature more than a conventional follow button: it connects a reader’s on-site choice to future visibility inside Google’s personalized and AI-driven search surfaces.
Google News on Android is making its daily audio briefings easier to tailor by topic, including broader categories and more specific subtopics. Google says the briefings include clear source attribution and links to the underlying full articles, giving listeners a route to read the original coverage.
Google News already lets users follow topics, locations and sources and manage what appears in their feeds. The updated briefings extend that personalization to a listening format, potentially making them more useful for people who want a quick update without browsing a traditional news feed.
Google says users have selected more than 600,000 unique sources, compared with more than 345,000 sources reported around the feature’s earlier launch.
That is a meaningful adoption signal, but it needs careful interpretation. The figure counts unique sources selected—not necessarily unique users, publisher implementations, completed visits or incremental traffic generated for publishers. Google’s public figure therefore shows that the feature is being used, but it does not reveal how much traffic any individual publication receives from it.
Google reports that links from a user’s Preferred Sources receive roughly twice the clicks of other cited links in the same AI response.
The result is encouraging for publishers, but it is not the same as an independent causal study. People who deliberately select a publication are likely to have a stronger interest in that source than ordinary searchers. They may also be more likely to click it regardless of the Preferred Sources label. The comparison therefore demonstrates a correlation in Google’s reported data, not a guaranteed traffic increase for every publisher.
There is also limited publisher-side visibility into the feature’s performance. Reporting has noted that publishers do not receive a complete dashboard showing how many users selected their site or how much traffic came from the relevant Search surfaces.
Preferred Sources is fundamentally a loyalty mechanism. A reader must first encounter a publication, decide to favor it and select the source. That makes the feature potentially valuable for established brands and repeat audiences, but less suited to replacing broad discovery traffic from readers who have not yet chosen a publisher.
The larger challenge is that AI-generated Search features can answer a question directly on the results page. A European Parliament research briefing says the available evidence points toward reduced publisher visits when AI-generated summaries appear. In that environment, a tool that increases clicks from users who have already chosen a source may improve the quality of some remaining referrals without restoring the overall volume of traditional search traffic.
Google’s own framing also highlights the trade-off: Preferred Sources helps readers find favored publications in Top Stories, AI Overviews and AI Mode, but those AI surfaces can still summarize information before a user visits the source.
Google’s AI-search controls add another layer to the decision. Publishers can opt out of appearing in or being used by AI Overviews and AI Mode while continuing to appear in ordinary Search results and the Discover feed.
That creates a difficult choice. Staying included preserves the possibility of being cited in AI results and benefiting from Preferred Sources, but also leaves the publisher exposed to a zero-click experience. Opting out limits participation in those AI surfaces and means giving up traffic or impressions from them, even though conventional Search visibility can remain.
The right decision will depend on each publisher’s priorities, audience behavior and ability to measure conversions beyond a single Search click. The available evidence does not support treating Preferred Sources as a proven replacement for lost referrals.
Google’s three updates point toward a more personalized content ecosystem:
Together, they may strengthen the relationship between readers and publications that already have trust or recognition. But they do not solve the broader structural problem created when Search answers more questions without sending users to the web. For now, the strongest conclusion is measured rather than promotional: Google is offering publishers new retention and visibility tools, while the scale of any traffic recovery remains unproven.
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Google is rolling out three personalization features: natural language controls for Discover, an embeddable Preferred Sources button, and topic based audio briefings in Google News on Android.
Google is rolling out three personalization features: natural language controls for Discover, an embeddable Preferred Sources button, and topic based audio briefings in Google News on Android. Discover users can describe what they want to see—or avoid—in ordinary language; Google may ask follow up questions, update the feed immediately and remember the requests.
Publishers can invite readers to add their site as a Preferred Source, while Google News audio briefings will offer more specific topics, source attribution and links to full articles.