91̽»¨

Up until now, finding a podcast on Spotify worked more or less the same way it always has. You searched for a name, browsed a chart or clicked something a friend recommended. If you didn’t already know what you were looking for, discovery was mostly luck – that changed this week.

Spotify has expanded its AI-powered Prompted Playlists feature to podcasts, letting Premium users in the UK, US, Canada, Australia, Ireland and Sweden type natural language requests to generate a custom listening queue from its catalogue of over five million shows. You can ask for “true crime about cold cases”, “tech interviews with startup founders” or “science and innovation with big discoveries” and Spotify will build you a queue.

The AI blends listening history, real-time trends and editorial charts to personalise results, and episodes include notes explaining why they were included.

Why This Is A Bigger Deal Than A Feature Update

The shift from keyword search to semantic understanding is a different game entirely.

Keyword search rewards the podcasts that are already well-known, because listeners have to know what to search for. Semantic search rewards the podcasts that best match what a listener actually wants, regardless of whether they have heard of the show before. For a catalogue of five million podcasts, the overwhelming majority of which are discovered by almost nobody, that is a structural change in how visibility works.

Spotify reports that over 34 million podcasts get discovered weekly on the platform – this number suggests the catalogue is more active than most people assume. Prompted Playlists adds a new discovery pathway that bypasses the chart and new-release mechanisms that have traditionally favoured established shows. A three-year-old episode from a niche show about cold-case investigations has the same chance of appearing in a listener’s queue as something that came out yesterday, if it best matches the prompt.

For independent creators and niche shows that have never cracked the charts, this opens up a path to audience growth they haven’t had before.

What This Means For Founders Using Podcasting As A Content Channel

A lot of founders and operators run podcasts as part of their content strategy because the format works well for B2B reach, thought leadership and building trust with a specific professional audience. If you’re in that camp, Spotify’s move deserves a closer look because it changes what discoverability actually requires.

Until now, podcast SEO has been primarily about metadata: titles, show descriptions, episode titles and transcripts. Those things still matter, but Prompted Playlists operates on semantic understanding rather than keyword matching. The underlying technology uses vector embeddings and natural language processing to match listener intent to content meaning. That means the question to ask about your podcast content is no longer just “does this have the right keywords?” but “does this clearly and consistently communicate what it’s about, who it’s for and what it feels like to listen to it?”

The good news for independent podcasters is that this levels the field against bigger, better-funded shows in a way that keyword-based discovery never did. In practice, this means that show descriptions, episode summaries and the language you use to talk about your content all become more important, because those are the inputs the AI reads when deciding whether your show matches a listener’s prompt.

The Broader Shift Happening In Audio Discovery

Spotify has been building toward this for podcasts dating back to at least 2022, when it moved from term-based search toward semantic matching for queries like “electric cars and climate impact.” Prompted Playlists for music launched earlier in 2026. The podcast expansion is the next logical step, and it signals where audio discovery is heading more broadly.

Audio is following the same playbook as every other content format. YouTube’s recommendation engine shifted creator strategy from thumbnail optimisation toward watch time and viewer satisfaction signals. Google’s shift toward semantic search changed what good SEO looked like for written content. Spotify is making the same move in audio, and the creators who understand the basic principles early tend to be the ones who benefit most from the transition.

The One Thing Worth Doing This Week

If you run a podcast, pull up your show description and your last ten episode summaries. Read them as if you were a listener trying to describe what you want in a natural language prompt. If the connection between what you write and what the show actually sounds like isn’t obvious, bridging the gap is a necessary step. Spotify’s AI is now reading your descriptions the same way a listener would describe a feeling, and the shows that communicate most clearly what they are will be the ones that show up.

Five million podcasts just got a new discovery layer. The shows that figure out the new logic early will find their audience faster than the ones waiting to be found.