Spotify is evolving its voice interface with generative AI, hinting at a more interactive experience ahead.
The company’s Chief Product and Technology Officer Gustav Söderström revealed during Spotify’s Q2 earnings call that users can now talk to Spotify in plain English to request songs. This interaction creates a new dataset that links phrases to songs, adding to Spotify’s existing playlist data.
Söderström explained: 
“You can think of it as us getting a new dataset. Spotify’s got this unique dataset from all of its playlists, which was really song-to-song — like which song goes well with another song, kind of the Amazon ‘people who bought this also bought that.’”
“And that’s completely new to us, and it’s a very, very valuable dataset that we are collecting very quickly.”
This dataset from voice commands is growing fast, setting the stage for “much more interactive” experiences.
“You can already write to Spotify, talk to Spotify. You’re just going to see that expand,” Söderström added.
Spotify’s AI will move beyond just predicting what you want to hear. It aims to “reason” over your listening history and voice interactions to offer smarter, multi-step responses.
The voice input comes from Spotify’s AI DJ feature launched in May 2025 for Premium users. It lets listeners talk to the DJ to change music, genres, or moods instantly.
Internally, Spotify is also leveraging generative AI to speed up product prototyping and improve finance operations.
Spotify reported 276 million paying subscribers in Q2, up 12% year-over-year, with 696 million monthly active users. But revenue shortfalls pushed the company to a loss, sending stock down 10%. CEO Daniel Ek expressed frustration with Spotify’s ad business during the call.
This push toward AI-driven voice interaction signals Spotify’s plan to make its streaming experience more conversational and responsive, not just predictive.
Read more on Spotify’s Q2 earnings and stock drop on weak guidance here.
