TL;DR: Netflix is testing a new "AI Discover" tab on its TV app that picks shows and movies based on your mood instead of making you scroll through another wall of thumbnails. It started rolling out Tuesday to English-language users in the US on supported TVs, and Netflix is openly calling it an experiment it might scrap if people don’t use it.
Netflix announced the feature on its own blog Tuesday, and TechCrunch was first to break down how it actually works. Users with access will see a new "Discover" button marked with a red gem icon and a "New" label in the top navigation bar on their TV. Tap it, and instead of the usual rows of posters, you get a set of curated themes built around what you’re actually in the mood for right now: something light, scare me, family movie night, quick watches, and a catch-all "more" option for anything else.
Pick a theme and Netflix serves up a shortlist of titles with clips attached, so you can preview before you commit. From there you can drill into a more specific theme to narrow things further, or hit "something different" to reset and start over. Each title also gets a "Tune Your Recommendations" button sitting right next to the usual Play and My List options, which lets you tell Netflix to save a title for later or rule it out of future suggestions entirely.
What’s actually new here
Netflix has been using algorithms to guess what you want to watch since long before anyone called it AI. Thumbs Up ratings, viewing history, and the My List queue have quietly shaped the homepage for years. What’s different about AI Discover is the framing: instead of Netflix deciding for you in the background, it’s asking you directly what mood you’re in and building the list around that answer in real time.
In its blog post, Netflix says the feature is powered by "a new intelligence layer" that responds to what users tell it and keeps learning from those signals. That’s about as much detail as the company is giving. There’s no mention of which model or technique sits behind it, no benchmark numbers, and no executive quote attached to the announcement, just a company blog post written in the collective voice. For a feature branded around AI, that’s a thin technical story, and it’s worth being honest about that rather than repeating the marketing copy as if it were a breakthrough.
What is clear is the problem Netflix is trying to solve. The company’s library has grown large enough that endless scrolling has become its own complaint among subscribers, and competitors have been chipping away at that exact pain point for a while. Netflix’s own language calls this tab a "test kitchen," a place to try recipes for discovery that might get folded into the main app later, or might just as easily get pulled if the data doesn’t back them up.
A limited, slow rollout by design
This is not a global launch. Netflix says AI Discover is starting with English-language users in the US who are watching on supported TVs, and that it plans to add more users "in the coming weeks." No specific TV platforms are named, no test group size has been disclosed, and there’s no timeline for other countries or languages. That’s a pretty narrow first step for a company with well over 300 million subscribers worldwide, and it suggests Netflix wants real usage data before it commits to anything bigger.
That caution makes sense given how much scrutiny any AI feature gets right now. Get the recommendations wrong, or make the interface more confusing than the current grid of rows, and the backlash writes itself. Netflix seems to be betting that a slow, reversible test beats a flashy keynote announcement, and given how often AI features from other companies have landed with a thud this year, that’s probably the smarter play.
It also fits a pattern Netflix has followed for most of its AI-adjacent moves. The company has generally avoided the kind of splashy, headline-chasing AI launches that define a lot of its tech peers, choosing instead to fold smaller improvements into the product and let the subscriber numbers do the talking. AI Discover looks like more of the same: a feature that can be measured by whether people actually tap into it and finish what they pick, not by how much attention it generates on launch day. If engagement holds up in the US test, expect a quiet expansion to more countries and languages rather than a big relaunch moment.
Part of a bigger shift in how people find things
Netflix isn’t alone in rethinking discovery as an AI problem rather than a search problem. Google’s AI Overviews now answer roughly 80% of searches directly on the results page, which means fewer people ever click through to the sites that used to do the finding for them. Reddit is running the same playbook from the other direction, testing ads inside its own search results partly because it’s losing referral traffic to those same AI-powered answers. The common thread across all three moves is that platforms are deciding users shouldn’t have to search at all if an AI system can just hand them the answer, or in Netflix’s case, the show.
That shift carries real tradeoffs. A mood-based tab that works well could genuinely cut down on the fifteen minutes some people spend picking something to watch, which is a legitimate problem worth solving. But it also hands Netflix another layer of control over what gets surfaced and what quietly disappears into the long tail, at a moment when the company is already juggling password-sharing crackdowns, ad-tier pushes, and a content library stuffed with licensed filler alongside its originals. An AI tab that nudges viewers toward titles Netflix wants watched, rather than titles viewers would have found anyway, is an easy line to cross without anyone noticing.
The bottom line
AI Discover is a sensible, low-risk experiment dressed up in AI branding that doesn’t actually tell us much about the AI underneath it. The mood-based themes solve a real and annoying problem, the rollout is cautious enough that Netflix clearly expects to iterate or kill it, and the lack of technical detail is a signal in itself: this reads more like a UX tweak with an AI label stapled on than a genuine model-driven leap. Worth watching if you’re a Netflix subscriber in the US test group. Not worth treating as evidence that Netflix has cracked recommendation AI in some new way, because the company hasn’t actually said it has.
The Teck Hustlers Editorial Team covers technology, business, finance and gaming news for readers who want the story behind the headline. Every article is researched, written and edited by our in-house team, checked against company announcements, filings and reputable reporting, and updated when new facts emerge. We focus on what each development means for founders, professionals, investors and everyday users. Have a tip or a correction? Reach us through our contact page.
