Spotify lets users prompt podcast playlists

Finding a great new podcast often feels like endless scrolling through categories, only to give up and replay an old favorite.

Spotify is betting that natural language prompts can fix that. The company just announced that its AI-powered Prompted Playlists feature, originally designed for music, now works for podcasts too.

Chat your way into a new queue

Currently in beta, the feature is rolling out to Spotify Premium subscribers in the U.S., Canada, the UK, Ireland, Australia, and Sweden, limited to English prompts for now.

To use it, you tap Create, choose Prompted Playlists, and type out exactly what you want to hear. Spotify offered an example prompt: Make a playlist of true crime podcasts I would be interested in. Add highly rated series I might have missed, especially ones full of twists and turns.

You can also set an update cadence for the playlist, choosing between daily, weekly on specific days, or keeping it static. Better yet, each episode added to the mix comes with a short explanation pointing out why it was picked for you.

Spotify's own culture team will also curate sample Prompted Playlists directly on the home screen, which users can grab and customize by editing the original text.

Unlocking the back catalog

Spotify first began testing Prompted Playlists in New Zealand late last year before rolling it out to more regions earlier this year. Adding spoken audio is the natural next step.

According to Spotify, listeners discover podcasts for the first time over 34 million times every week. Lizzy Hale, Spotify's global head of podcast editorial, noted that the tool helps creators surface back-catalog material alongside new episodes to listeners who are actively telling the app what they want.

The push comes alongside aggressive efforts to woo podcasters. Earlier this year, Spotify slashed the monetization barrier for video podcasts down to three published episodes, 2,000 consumption hours, and 1,000 engaged listeners over a 30-day window. Co-CEO Alex Norström recently noted during an earnings call that the platform hosts over 530,000 video shows.

Why it matters

For podcast listeners, this cuts down the steep friction of finding something worth forty minutes of your day. Unlike music, where skipping a track costs three seconds, trying a bad podcast episode wastes meaningful time. Prompting by vibe, plot tropes, or hyper-specific topics lowers that trial hurdle, provided you are a paying subscriber in one of the launch countries.

For creators and the wider industry, this could finally breathe life into archive libraries. Traditional podcast apps rank shows by overall popularity and recent release dates, leaving older episodes to gather digital dust. If natural language discovery works as advertised, a great evergreen episode produced three years ago could suddenly find an audience on demand, putting pressure on competitors like Apple Podcasts to offer similar semantic curation.

We think Spotify's cleverest detail here is the short note explaining each pick. Recommendation engines often feel opaque, and justifying the inclusion makes users far more likely to give an unfamiliar host a shot. The risk, of course, is prompt hallucination or weak indexing: if the AI serves filler episodes masquerading as high-stakes thrillers, listeners will quickly abandon the feature and head back to word-of-mouth recommendations.