Google knows a lot of things about its users. Now, the company wants to put some of that knowledge to use to improve music recommendations: Google relaunched its Google Play Music service with a focus on context-relevant content recommendations Monday. The new version will roll out to all users across Android, iOS and the web this week.

“Google Play Music uses machine learning to figure out what music you like and then mixes in signals like location, activity, and the weather along with hand-picked playlists to personalize music for wherever you are and whenever you want tunes,” explained lead product manager Elias Roman in a blog post Monday.

Users who opt into using Google data to power their music recommendations will get song suggestions based on the place they’re listening, their current activity and even the weather. “Your workout music is front and center as you walk into the gym, a sunset soundtrack appears just as the sky goes pink, and tunes for focusing turn up at the library,” Roman wrote.

In addition to the new content recommendations, Google Play Music is also getting a new look, as well as improved offline listening: Instead of asking users to download their favorite albums beforehand, Google Play Music’s new mobile apps do this automatically based on past listening — a feature that’s similar to the way Pandora’s new $5 tier caches music for offline listening.

Google has yet to release subscriber numbers for Google Play Music, but it’s widely assumed that the service trails more popular competitors like Spotify and Apple Music by a wide margin. That’s in part because there has been little to differentiate the service from its bigger competitors. Using context-relevant content recommendations could be a first step to more clearly define what Google Play Music is all about.