Google is rolling out a new feature to the Google app that allows users to customize their Discover feed using a chatbot-style interface, enabling them to describe preferred content types and have the AI remember those preferences for future visits, according to reporting by The Verge.
Rewriting the Algorithm Through Conversational Prompts
Google is now changing that interaction model. In the coming days, an update hitting the Google app introduces direct natural-language control to the Discover feed.
Tapping the standard three-dot menu on any card within the Discover feed now exposes the customization entry point. According to demonstration videos released by Google, selecting this option launches an interactive dialogue box. Here, users type out prompt-based parameters detailing specific topics, angles, or recurring themes they wish to prioritize.
Under the Hood of LLM-Driven Feed Personalization
The chatbot interface does not just ingest the prompt and vanish. It actively confirms the user’s choices, laying out an explicit summary of the content categories it will prioritize moving forward. If the model misinterprets a nuance or misses a sub-topic, the interface provides an iterative feedback loop, allowing users to inject supplementary details. This persistent memory architecture ensures that subsequent visits to the Discover feed reflect these conversational adjustments over time.
Ecosystem Dynamics and User Control
The 30-Second Verdict
- What is changing: The Google Discover feed is gaining a chatbot-driven customization feature accessed via the three-dot menu.
- How it works: Users type natural-language descriptions of their preferred content, and the AI updates and remembers their ranking preferences.
- Timeline: Rolling out to the Google app in the coming days.
