Google Discover will soon let you tune your feed with an AI chatbot

Google is adding a chatbot-style customization tool to Discover, allowing users to describe the kind of stories they want to see and have AI reshape the feed accordingly. The feature will arrive in the Google app in the coming days and matters because it gives people a more direct way to steer one of Google’s biggest recommendation surfaces.

Instead of relying only on passive signals and broad preferences, Discover will soon accept natural-language instructions. Google says the system will remember those choices for future sessions, making the feed feel more personalized over time rather than resetting every time a user opens it.

What Google is changing in Discover

Google Discover, the content feed that appears in the Google app and on some Android surfaces, is getting a new AI-powered control layer. The company is introducing a prompt-based interface that lets users type or speak what they want more of, then adjusts the feed based on that instruction.

The change is designed to make Discover feel less like a black box. Right now, the feed is largely shaped by activity across Google Search and other Google products. With the new option, users will be able to give the system explicit guidance, such as wanting more sports analysis, local news, or coverage from a specific topic area.

Google says the update will roll out soon through the Google app, where users will find it inside the three-dot menu attached to Discover cards and sections. From there, the experience is meant to work like a lightweight chatbot: the user describes preferences, the system responds with what it thinks it understood, and then the user can refine the request if needed.

According to Google’s demonstration, the feature is built to confirm user choices, explain how it plans to change the feed, and accept follow-up instructions if the first interpretation is not quite right.

How the new AI feed tuning works

The new Discover control uses natural language rather than toggles or fixed topic lists. That matters because users often know what they want in broad terms, but not necessarily in the exact settings language that older personalization tools require.

Google’s interface is intended to remove that friction. A user can describe the type of content they want, see a summary of how Discover will adapt, and then press a refresh button to apply the changes. Google says those preferences will be remembered, so the feed can continue adapting in subsequent visits.

How users will access the feature

Google says the option will live in the three-dot overflow menu on Discover. That keeps the tool close to the feed itself, rather than hiding it in a separate settings area that many people never open.

That placement is important because Discover is built around rapid, on-the-go consumption. By putting the customization feature inside the feed, Google is encouraging users to make small adjustments in context, right when they notice something they want more or less of.

What the chatbot-style flow looks like

Google’s demo suggests the interaction is meant to be conversational, even if the underlying function is simple personalization. The user submits a request, Google’s system restates what it understood, and the user can add clarification before confirming the change.

That kind of loop is increasingly common across consumer AI products because it lowers the barrier to entry. Instead of learning settings labels or browsing menus, people can write the result they want in plain English.

Why this matters for Google and for users

The update is part of a larger shift in how major platforms manage recommendations. For years, feeds have been driven by opaque ranking systems that infer user interests from behavior. AI now gives companies a way to add a more obvious layer of user control on top of those systems.

For users, the upside is clearer personalization. People are increasingly frustrated by feeds that over-index on one topic, repeat the same publishers, or take too long to recover after a single click on something they do not really want. A prompt-based tool can be faster and easier than hunting through settings.

For Google, the feature may also help Discover feel more responsive at a time when the company is integrating AI across more of its consumer products. Google has been pushing AI-driven interfaces in search, news, and other apps, and this update extends that strategy into a mainstream recommendation feed.

Why feeds are becoming more conversational

Platforms are leaning into conversational controls because they mirror the way people already think about preferences. A user is more likely to say “show me fewer celebrity stories and more local politics” than to navigate a dozen filters.

That is why similar features have appeared elsewhere. Google notes that other services, including YouTube, Instagram, Bluesky, and X, have also introduced AI-assisted ways to influence what users see. The broader industry trend suggests that feed personalization is moving from implicit ranking toward explicit, user-directed tuning.

What else Google announced

The Discover change was not the only product update Google highlighted. The company also said it is adding personalization options for daily audio briefings in the Google News app on Android.

That update extends the same idea into another format: rather than only changing the reading experience, Google wants to shape how people listen to news summaries as well. It is another signal that the company sees personalization as a major part of its media strategy across apps.

Google also updated Preferred Sources, a feature that lets people see more results from selected publishers in Search’s top stories area, as well as in AI Overviews and AI Mode. The company says publishers will now be able to place an interactive Preferred Sources button on their own websites so readers can add them more quickly.

How Preferred Sources changes publisher discovery

The new button is meant to reduce friction. Instead of forcing readers to leave an article, visit Google’s settings, and manually add a source, publishers can now provide a shortcut on-site.

That could matter for outlets that rely on loyal readership. The easier it is for users to indicate a preferred publication, the more likely a source is to remain visible across Google’s news surfaces.

How does this fit into the broader AI feed race?

It fits into a larger competition among tech companies to make recommendation feeds more editable, explainable, and responsive. Social and search platforms increasingly view AI as the layer that can translate a user’s vague intent into concrete feed changes.

That push is partly about user satisfaction and partly about retention. If people can fine-tune feeds quickly, they may spend more time inside an app because the content feels better aligned with their interests. In that sense, personalization is both a convenience feature and a strategic product lever.

Google’s version is notable because Discover already occupies an important position in the mobile experience. It sits at the intersection of search, news, and recommended content, which gives Google a strong incentive to make the feed feel both useful and controllable.

What users can expect next

Google says the feature will begin appearing in the coming days, suggesting a staged rollout rather than a global, same-day launch. That is typical for consumer product changes, especially when they involve AI-driven recommendations that may need monitoring and adjustment.

Users should expect the customization tool to show up in the Google app’s Discover feed menu first. Once it arrives, the main workflow will likely involve a prompt, a confirmation step, and a refreshed feed that reflects the new preferences.

Because the system is designed to remember choices, the impact may not always be immediate and dramatic. The real value could be cumulative: over time, Discover may become more in line with a user’s stated interests and less dependent on guesses made from past behavior alone.

Timeline of the rollout

Here is a quick breakdown of the changes Google described.

Why Google’s move is strategically important

Google is making a calculated bet that users want more agency over algorithmic feeds without having to abandon automation altogether. That balance is central to the company’s current product direction: use AI to reduce effort, but keep the experience familiar and fast.

The Discover update also reflects an important design philosophy. Rather than building a separate assistant for feed curation, Google is embedding AI into an existing habit loop. People already check Discover to see what is new; now they will be able to shape that stream in the same place.

In a crowded app ecosystem where attention is scarce, small utility gains can have outsized effects. If this feature makes Discover feel more relevant, it may strengthen one of Google’s most visible recommendation products while also giving users a rare sense of control over what the algorithm serves up.

What this means for publishers and platforms

Publishers may benefit if the new Preferred Sources controls make it simpler for readers to indicate loyalty to specific outlets. At the same time, the Discover AI tuning tool could make content distribution even more dependent on explicit user preference, not just broad popularity signals.

That could reward outlets with strong brand recognition and niche audiences, while making it harder for generic content to break through unless it closely matches user intent. It also raises the stakes for publishers competing to stay visible in Google’s expanding AI-driven surfaces.

For platforms more broadly, the update is another sign that the next generation of feeds may look less like a static recommendation engine and more like a guided conversation. The question is no longer whether AI can predict what people might like. It is whether AI can help people tell the platform what they want, in their own words, and make the result feel better than the old model.