Netflix is rolling out a new AI-powered Discover experience that replaces the familiar scroll-and-guess routine with something closer to a conversation. Instead of hunting for a title or filtering by genre, subscribers can now describe a mood, "something funny and upbeat," "a cozy night-in movie," or “cheer me up”, and let the platform surface what fits. The feature is built on a version of OpenAI's technology that Netflix has trained on its own catalogue, and it sits at the center of the company's most sweeping interface redesign in years.
Introduced by Chief Product Officer Eunice Kim and Chief Technology Officer Elizabeth Stone at a company tech and product event, the tool began as an opt-in iOS beta in Australia and New Zealand before expanding to the United States and other markets. Netflix's framing is deliberately human: the goal, executives said, is to make searching feel less like filing a query and more like asking a friend what to put on.
BY THE NUMBERS
~110 hrs spent per year by the average U.S. streamer just deciding what to watch | 54.9% of viewers give up watching entirely due to decision fatigue | 325M+ Netflix global subscribers as of early 2026 |
The problem Netflix is actually solving
The pitch lands because the pain is real and measurable. A UserTesting survey of 2,000 American subscribers found the average viewer burns roughly 110 hours a year, nearly five full days, scrolling in search of something to watch. Researchers have a name for the loop: "Netflix Syndrome," in which people spend longer choosing than viewing and often give up or default to a rerun.
More recent data suggests the fatigue is deepening. A QuestionPro study published in October 2026 reported that 61.3% of Americans feel overwhelmed by the sheer number of options, 54.9% have abandoned a viewing session entirely because they couldn't decide, and 77.4% retreat to familiar shows rather than risk a new one. A separate Bango survey of 2,500 U.S. consumers found that 30% say choosing what to watch frequently takes longer than watching it, a figure that climbs to 46% among Gen Z and Millennial respondents.
Netflix is not exempt from the frustration it is trying to fix: in the Bango data, 36% of its own users said selection often outlasts satisfaction. For a library this large, discovery isn't a feature, it's the bottleneck.
How the Discover tab works
Talk to it like a person
The core shift is from keywords to natural language. Rather than matching exact titles, actors, or genres, the system parses intent. Type "martial arts musicals" or "I need a cozy night-in movie" and it returns tailored picks drawn strictly from Netflix's own catalogue, a design choice the company says keeps recommendations grounded and reduces the risk of AI hallucinating titles that don't exist on the service.
Recommendations that shift with your mood
The redesign extends beyond the search box. Netflix says the homepage now responds to a viewer's "moods and interests in the moment," with larger tiles, fewer cluttered rows, and at-a-glance context explaining why a given title is being suggested. Once a result appears, users can start watching, add it to My List, or share it, turning discovery into something more interactive than a static grid.
A TikTok-style feed, too
On mobile, Netflix is pairing the Discover tool with a vertical, swipeable feed of clips and trailers, an acknowledgment that short-form browsing is how many people now stumble onto their next obsession. Kim noted that members already love browsing clips to find something new; the feed simply formalizes that behaviour.
Why now: the engagement math
The timing isn't accidental. Netflix crossed 325 million subscribers in early 2026 and logged roughly 97 billion viewing hours in the first half of the year, up about 2%, but much of that growth came from adding members, not from existing ones watching more. Analysts at Jefferies estimated that average monthly viewing per member slipped from about 66 hours in the first half of 2023 to roughly 49 hours in the first half of 2026.
In other words, the company is adding accounts while each account leans in a little less. Churn remains low, around 2%, and U.S. households still give Netflix more daily time than any rival (about 3.5 hours a day in June 2026, per Comscore). But in an increasingly ad-supported business, time spent is the metric that pays. A discovery engine that converts aimless scrolling into actual play is, in that light, a direct lever on revenue.
If the old problem was "too much content, too little time," Netflix's bet is that the fix isn't more content, it's a smarter way in.
OpenAI under the hood, but on Netflix's terms
The conversational layer is powered by OpenAI's ChatGPT technology, but Netflix has stressed that its own team trains and tunes the models to fit member needs and the specifics of its catalogue. Grounding the system in Netflix's library rather than the open web gives the company tighter control over behaviour, privacy, and training data, and limits the odds of the assistant confidently recommending a film the service doesn't carry.
It is also part of a broader AI push. Netflix has said creative partners used generative-AI workflows across roughly 300 titles in 2026, concentrated in post-production, and the company is experimenting with AI to localize title cards into subscribers' preferred languages. Discovery is the most visible piece of that strategy because it's the part subscribers actually touch.
The catch: smarter search, same library
Not everyone is convinced a better front door solves the deeper issue. The most common refrain among early reactions is blunt: a more intuitive way to search still only surfaces what's already on Netflix, and many users' real complaint is that nothing in the catalogue grabs them in the first place. Bloomberg has also reported that viewership for Netflix originals tends to fall off sharply after a first season, a content problem no search box can fix.
There are open questions, too, about how an AI that optimizes for engagement shapes what people watch, and how much mood data subscribers are comfortable handing over. Among households with children, surveys show rising unease about recommendation algorithms steering young viewers. For now, Netflix is positioning Discover as a convenience, not a gatekeeper, subscribers can still ignore it and search the old way.
What happens next
Netflix plans a gradual, market-by-market rollout, with the iOS search experience arriving first in select regions and the redesigned TV interface reaching users worldwide over the coming weeks and months. Whether mood-based discovery meaningfully lifts engagement, rather than simply repackaging the same catalogue in friendlier language, will be the metric to watch when Netflix next reports results.
The stakes extend past one company. If describing a feeling becomes the default way hundreds of millions of people choose entertainment, Netflix won't just have redesigned its app; it will have nudged the entire industry toward conversation as the new interface for discovery. The scroll, in that future, becomes the exception rather than the rule.
Olivia Bennett is a technology writer at TechJox with over 8 years of experience creating detailed how-to guides, software tutorials, troubleshooting articles, and step-by-step walkthroughs. She specializes in Windows, Android, iOS, AI tools, software applications, browser settings, and everyday tech solutions. Her expertise includes simplifying technical processes, resolving common software issues, and explaining digital tools through practical, beginner-friendly instructions. Olivia focuses on making technology easier to navigate, helping readers solve problems confidently and get the most out of their devices and applications.
Discussion (0)
You need a free TechJox account to join the discussion and report scams you spot.