Netflix just rolled out a new generative AI-powered search tool that lets you use natural language to find what you’re in the mood to watch—no keyword hacking required.

But this isn’t just a search bar upgrade—it’s Netflix quietly redesigning how we discover content, one AI query at a time.

The new feature, announced May 7, is built on in-house large language models (LLMs) trained specifically on Netflix’s catalog and user data.

It lets you type prompts like “show me Korean dramas with strong lead character” or “what’s a good TV show to watch when I want to binge-watch” and get personalized results, instantly.

Think of it less like browsing and more like briefing your own entertainment assistant.

For now, it’s available only to a limited number of English-speaking users in the US and Canada using TVs, with mobile and web support planned later this year.

That slow rollout makes sense—it gives Netflix time to gather feedback and refine how the AI handles complex requests before scaling it across devices.

This is part of a broader trend: streaming platforms are under pressure to cut through their own noise.

As libraries balloon and competition stacks up, a search function that actually understands what you want—without making you dig—isn’t just nice to have, it’s essential UX.

Netflix’s design here is subtle but intentional, leaning into conversational prompts and cutting down on click fatigue. It’s less about menus and more about moods.

While Netflix hasn’t revealed the model architecture powering the tool, it’s likely a custom-tuned version built in-house, not off-the-shelf from OpenAI or Anthropic.

That gives them tighter control over privacy, training data, and overall behavior.

And by grounding the model in its own media catalog, Netflix reduces the risk of AI hallucinations—something that’s still a known issue for broader gen AI tools.

This update also hints at something deeper: Netflix is using generative AI not just for content creation (like its experiments in anime storyboarding), but for user interface design.

This is AI as infrastructure—making the platform feel smarter, smoother, and more aligned with how people actually talk.

The bottom line is Netflix doesn’t want you scrolling, it want you watching.

It wants you saying what you feel and watching what fits—fast.

With AI stepping in as the go-between, the line between algorithm and assistant just got a lot thinner.