Netflix Debuts AI Discover: Mood-Based Recommendations Arrive on TVs

Netflix has begun rolling out AI Discover, a new tab on its TV app that lets subscribers describe what they feel like watching and receive recommendations generated in real time. The feature, announced on 3 November 2026, is the first user-facing Netflix application the company is explicitly branding as AI-powered, and it marks a shift away from endless scrolling through rows of artwork toward a more conversational discovery experience.

• Mood-first discovery: Users pick from themes such as “Something Light,” “Scare Me,” “Family Movie Night” and “Quick Watches,” each previewed with a clip from a show or movie.
• Limited initial rollout: AI Discover is available first in the United States to members watching Netflix in English on supported TVs, with a gradual expansion planned over the coming weeks.
• Built-in feedback loop: Every title carries a “Tune Your Recommendations” button beside “Play” and “My List,” letting viewers keep a suggestion or ask Netflix not to recommend it again.

How the AI Discover experience works

The first experience inside AI Discover is built around themes rather than a search box. After opening the tab, members see a carousel of themes generated around what Netflix believes matters to them right now. Those themes can reflect a desired mood, the people they are watching with, how much time they have, what is new or trending, or what they have been watching lately.

Selecting a theme brings up a set of matching titles. More specific themes then surface so users can keep narrowing the selection, or they can choose “Something Different” for a fresh set. Each recommendation includes the usual synopsis, metadata, trailer and box art, so the experience remains familiar once a title is chosen.

Netflix says the feature is powered by a new intelligence layer it has been building that responds to what members say in real time and keeps learning. The company has not named a specific external AI model, and it describes AI Discover as testing only a subset of what that layer can do.

“AI Discover is powered by a new intelligence layer we’ve been building that responds to what members tell us in real time and keeps learning,” a Netflix spokesperson said. “We can understand what [subscribers] are in the mood for right now, respond to what they ask, and open up new paths to discover beyond the ones that we present are predefined.”

The tab appears in the top navigation bar with a red gem icon and a “New” label. It is not available on kids’ profiles, and members cannot sign up for it directly; access is allocated as part of the test. Those included cannot opt out or remove the tab, though they can simply ignore it.

A test kitchen for discovery

Netflix is positioning AI Discover as a work in progress rather than a finished product. The company calls the tab its “test kitchen,” where it will try out new recipes for discovery and use member feedback to decide which features to keep, improve or serve up next. Some of those ideas may eventually make their way into the main Netflix experience.

“Think of AI Discover as our test kitchen: we’re trying out new recipes for discovery, and your use and feedback help us decide which ones to keep, improve or serve up next,” Netflix wrote in a blog post.

The launch comes as Netflix looks to make its growing library easier to navigate and to lift engagement. The company has long used machine learning to personalize recommendations, but AI Discover is the first application it is branding as AI-driven. Netflix has also adopted AI elsewhere in its business, including generative AI in the production of roughly 300 titles and its acquisition of InterPositive, the AI tools company founded by Ben Affleck, for $587 million.

Netflix says feedback from AI Discover will not be used to inform programming decisions. The feature is purely about recommendations, in the same way its existing recommendation systems work today.

What the rollout signals

The limited US launch on supported TVs gives Netflix room to observe how members actually use mood-based discovery before committing the feature more broadly. The company has not said how many people are included in the test, nor which TV platforms are supported beyond the general description of compatible televisions.

The strategic bet is clear: reduce the friction of choosing something to watch by letting viewers state their intent directly. Whether that produces better decisions or simply adds another algorithmic layer over an already crowded catalog will depend on how the intelligence layer learns from the “Tune Your Recommendations” signals. For now, AI Discover remains an experiment with a visible label, a narrow audience and a promise that user behavior will shape what comes next.