SAN FRANCISCO – You can now talk to Spotify. Not at it, but with it, back and forth, in natural language, about what you want to hear next or what you heard three years ago when you were in a very specific mood. The company launched a conversational AI assistant on Monday for Premium subscribers in the United States, Ireland, and Sweden, marking the most consequential change to how people find music on the platform since the AI DJ feature arrived in 2023.
The feature accepts text or voice input from the Home and Now Playing screens on iOS and Android. Users can ask the assistant to build playlists, explain what they are currently listening to, pull up specific artists, or ask questions about their own listening history. “When did I last play this song?” is now a question Spotify can answer. The conversation continues in context. Follow-up questions work. Spotify calls it “Talk to Spotify,” and the beta is available now to Premium members who are at least 18 years old.
Spotify confirmed to TechCrunch that the assistant uses “a mix of its own AI technology and models from multiple providers, based on whatever is best for the task.” The company declined to name those providers. That vagueness matters. How well conversational music discovery works depends heavily on the underlying model’s ability to hold context across turns, parse ambiguous language, and map natural descriptions onto tracks and artists in a catalog of more than 100 million songs. Whether it does that reliably is what the beta is meant to find out.
The assistant does not replace Spotify’s existing AI tools. The AI DJ, which has been in the app since 2023 and generates narrated personalized mixes based on listening history, continues running separately. The AI playlist-building tool, which already accepts text prompts, remains its own function. The new assistant sits on top of these as a conversational entry point that can invoke them, rather than replacing either one. Spotify described it as the interface users will eventually reach for first, before the search bar.
Every major platform company is currently building something that claims to sit between users and content through a conversational interface. Apple has a native AI agent in Safari capable of acting on files stored on a user’s computer. OpenAI is building a home speaker designed to learn its owner’s email and daily habits before being asked to. Meta’s own AI agent push has not accelerated at the pace Zuckerberg expected, despite a $145 billion infrastructure commitment. In music, Apple Music and Amazon Music both offer voice-enabled features through Siri and Alexa, but neither was designed around conversational music discovery. Spotify’s bet is that a music-native assistant with deep listener data does this better than a general-purpose voice system bolted onto a catalog.
The service launched in three markets partly by design and partly by constraint. Spotify’s European operations are headquartered in Ireland, and Sweden is the company’s home market, making both natural testing grounds. The English-only requirement and the 18-and-older restriction suggest the company is treating this genuinely as a beta rather than a staged global rollout under a beta label. No expansion timeline was announced. No pricing changes were tied to the feature.
What the official announcement does not address is how the new assistant compares to the AI DJ at its core task: surfacing music a user did not know they would like. The DJ already has access to the same listening history the assistant references. Its algorithmic model has years of behavioral data behind it. The conversational layer adds addressability, the ability to say precisely what you want rather than waiting for the algorithm to guess it. Whether that produces better discovery outcomes, or just feels like it does, is the question the beta is quietly running.
The business logic behind this launch is not subtle. Spotify ended 2025 with roughly 700 million monthly active users but has been deploying AI features specifically as a retention instrument rather than a growth one. A listener who finds the right song through a conversation stays longer than one who abandons a playlist two songs in. Earlier AI integrations positioned Spotify as a platform where third-party AI systems could connect and discover, not just a passive catalog. This is the company’s most direct attempt to own the conversational discovery layer itself, rather than hand it to Anthropic or OpenAI to run.
The harder question is whether the people most likely to appreciate a music chatbot are already well-served by what Spotify has. Power users with specific tastes already know how to use search, filter by mood, and skip what the algorithm gets wrong. Casual listeners who stream whatever is popular may not reach for a conversational feature at all. The users most likely to benefit are somewhere in between: people who know what kind of thing they want but cannot express it in a keyword search. That is a real use case. It is also a narrower one than Spotify’s announcement suggests.
What Spotify launched on Monday is a conversational front-end layered onto algorithmic systems that already exist. The company is confident enough to ship it and measured enough not to call it anything more than a beta. Whether talking to your music app changes what you listen to, how long you stay in it, or how much you trust what it recommends is not answerable from the feature’s first day. The data will not exist until the beta is considerably larger. That is precisely why it is a beta.
