Spotify is handing listeners a microphone for their algorithm — not literally, but close. The company’s new beta, called Prompted Playlists, lets Premium users write what they want to hear in plain English and then uses that instruction plus the entirety of their Spotify history to build and refresh a custom playlist.
What it does
Rolling out to Premium users in New Zealand starting December 11 (English only, for now), Prompted Playlists asks you to type a prompt — anything from a short phrase to a multi-sentence brief — and then curates songs that match both that brief and your listening fingerprint. You can schedule the playlist to refresh on a cadence (daily, weekly), tweak the prompt, or start over. Spotify says each song will include a short explanation of why it was chosen, and there’s an “Ideas” tab to nudge you if the blank page feels intimidating. For Spotify’s own explanation, see their announcement on the Spotify newsroom.
It’s presented as an evolution of Discover Weekly and other personalization tools: instead of passively receiving whatever the algorithm thinks you want, you get to set the guardrails and the creative brief. Examples Spotify offered include things like “music from my top artists from the last five years, but deep cuts I haven’t heard yet” or “high-energy pop and hip-hop for a 30-minute 5K run that keeps a steady pace before easing into relaxing songs for a cool-down.”
Why this matters (and where it might trip)
On the surface, Prompted Playlists solves a familiar frustration: the algorithm is powerful but opaque. Letting users direct it with natural language turns that black box into a co-pilot you can steer. For creators, that could mean smarter discovery — songs matching an exact mood or use case are more likely to reach listeners searching for those contexts.
But there are trade-offs. The feature leans heavily on long-term listening data — Spotify says it taps the “full arc” of your taste — which raises predictable privacy and control questions. As AI systems grow more embedded in everyday apps, conversations about what data these models can access and how they use it have intensified (see ongoing debates about AI tools that peer into personal inboxes and cloud storage in projects like Google’s Gemini research) — those are the same kinds of issues that will shadow any feature that ties deep personal history to generative recommendations. For broader AI privacy and integration concerns, read more about Gemini Deep Research and related privacy questions.
There’s also a cultural angle: platforms are now selling algorithm tuning as an experience — a flip of the old promise that smart systems would just figure out what you want. Instagram and other apps have begun adding controls that let users steer recommendation engines, and Google’s experiments around agentic AI show the same impulse to give people more direct levers over automated systems. Spotify’s move fits squarely into that trend; if you want more context on how companies are testing user-facing AI controls, see reporting on Google’s AI Mode agentic features.
Small details that change the experience
A few practical bits set Prompted Playlists apart from prior AI playlists:
- Prompts can be long and specific; Spotify says the system factors in world knowledge so you can reference current events or cultural touchpoints.
- Playlists use your entire listening history from day one, not just recent plays. That means a teenage obsession could resurface in a perfectly reasonable way — or creep into a mood where you didn’t want it.
- The song-level explanations could help users learn how to phrase better prompts, turning the interaction into a feedback loop rather than a single shot.
Try these two starters: “30-minute upbeat indie and electro for focused work — include artists I’ve saved but rarely played” or “soft ballads from the 2010s for late-night reading, prioritize underrated singers.” Those prompts show how you can combine tempo, era, emotional tone, and your own listening behavior.
Artist and industry angles
If the feature scales, it could be meaningful for artists and labels looking to surface catalog tracks or niche releases to targeted listeners. But as with any algorithmic discovery tool, the winners will be those who understand how the system matches language, metadata, and actual audio features. That may push some teams to rethink metadata, release messaging, and how they engage with playlists.
Should you try it?
If you’re the sort who tweaks playlists obsessively, Prompted Playlists is going to feel like a fast pass to exactly what you want. If you care about how much of your history a platform uses, this is a moment to check privacy settings and read Spotify’s beta notes. Either way, it’s another sign that personalization is becoming interactive rather than passive.
And if you do decide to test out a new running mix or late-night soundtrack, pair the experience with a decent set of buds — say, AirPods — so the transitions and production shine.
Spotify is calling this the start of a broader shift toward giving listeners more control, and the company plans to iterate as the beta expands. Expect the feature to evolve as Spotify balances personalization, discoverability, and the thorny questions that come with making algorithms more conversational and more in the hands of everyday users.