Spotify Is Letting Readers Ask AI What They Should Read Next — And Book Discovery Is Starting to Change
By Wendy Keir | EmpowerAi™
Spotify has introduced a new way for people to discover audiobooks.
Instead of searching by author, title or category, listeners can now describe what they are looking for in ordinary language and ask Spotify to create a personalised audiobook playlist.
Someone could ask for books with acclaimed screen adaptations from the past three years. They could ask for one audiobook a month matched to the season and limited to a particular listening length. Spotify combines the request with the listener's own preferences and what is currently happening across books to create the recommendations. The playlists can then refresh daily or weekly.
That is interesting enough on its own.
What caught my attention is that Spotify has also started allowing some listeners to talk to Spotify about the books they are listening to.
We are beginning to move from book search towards book conversation.
Readers have traditionally needed to know what to look for
Book discovery has always involved a certain amount of luck.
Someone recommends something.
You hear an author on a podcast.
Amazon shows you a title because you bought something similar.
You walk into a bookshop and notice a cover.
Search helps, although search works best when you already have some idea what you want.
If I type leadership books into a search engine, I have already decided that leadership is the subject and that a book is the form of solution I want.
Conversational discovery starts earlier than that.
A reader might say that they have recently become responsible for a small team, they are struggling with difficult conversations and they want something practical rather than academic.
That description contains considerably more information than the words leadership books.
Spotify's Prompted Playlist for audiobooks begins to make use of that difference. The reader describes the kind of experience they want and the system works out which books may fit.
That changes the route into the book.
This could matter enormously for authors who do not already have a large audience
Discovery has always favoured familiarity to some degree.
Readers recognise an author.
They recognise a publisher.
They see a book repeatedly until eventually they decide to investigate it.
Conversational recommendation has the potential to give relevance a larger role.
A relatively unknown author may have written precisely the book someone needs, although the reader would never have known what title to type into a search box.
If AI can understand the underlying need and connect it with the right book, the size of the author's existing audience becomes only one signal among many.
I would be cautious about claiming that this creates a level playing field. Recommendation systems still have algorithms, commercial relationships, behavioural data and all sorts of other influences underneath them.
Even so, the mechanism is different.
The reader can increasingly describe the problem rather than identify the product.
That is a substantial shift.
Spotify is also beginning to make the listening experience conversational
The second development may eventually be even more important.
Spotify says eligible users in the US, Ireland and Sweden can now type or speak to Spotify and have a back-and-forth conversation about music, podcasts and audiobooks. They can ask for recommendations, follow up on previous requests or ask questions connected with something they are currently listening to.
For example, Spotify gives the example of somebody asking where the Cyclops first appears in The Odyssey.
That might sound like a fairly ordinary AI feature.
In publishing terms, I think it points towards something bigger.
The book is no longer necessarily a closed object that the reader consumes and then leaves in order to find an explanation.
The conversation can begin while they are still inside the experience.
This is where it starts becoming particularly interesting for nonfiction
Imagine somebody listening to a business book.
The author introduces a framework.
The listener understands most of it but is uncertain about one part, so they stop the audiobook and ask a question.
At the moment, a general conversational system may be able to explain what happened in the book or retrieve factual information about it.
A governed AI Book Companion™ can do something rather different.
It can help the reader explore the question inside the author's own methodology.
That distinction matters.
If the reader asks how a particular framework applies to their situation, the Companion can work with the author's definitions, approved material, examples and boundaries rather than simply producing an answer based on general knowledge.
Spotify's new features are not doing that, and I would not suggest they are.
What they demonstrate is that readers are increasingly going to expect conversation to sit much closer to the act of reading and listening.
That is the part I think authors should notice.
Discovery and interaction may begin moving together
Something else happens when you put these pieces together.
The same environment that recommends the book is beginning to support conversation about the book.
Traditionally those were separate stages.
You discovered the book through a bookshop, search engine, recommendation or advert.
You bought it somewhere else.
You read it.
If you wanted to explore an idea afterwards, you might visit the author's website, join a programme, search for an interview or simply move on.
Digital platforms have already compressed some of that journey.
AI may compress it much further.
The system could help somebody articulate what they need, discover a relevant book and then remain present while they work through what the book means.
That starts to change what a book can sit inside.
Authors may need to think beyond metadata
For years, book discoverability has involved metadata.
Title.
Subtitle.
Categories.
Keywords.
Description.
Those things will continue to matter because platforms still need structured information.
Conversational discovery introduces another consideration.
Can a system understand what problem this book actually helps somebody solve?
For nonfiction authors in particular, that may become increasingly important.
An author may describe their work as leadership development, although the actual reader problem is much more specific. Perhaps the book helps somebody who has recently become a manager and feels uncomfortable asserting authority. Perhaps it helps a founder whose business has outgrown the way they have always made decisions.
Those distinctions are difficult to capture in a keyword.
They are much easier to express in conversation.
If discovery systems increasingly work from reader intent, clearly articulating those distinctions becomes part of discoverability.
There is still a trust question underneath all of this
Recommendation is never completely neutral.
Spotify's system knows something about the listener's history and preferences. It also uses what is currently trending in the literary world when building prompted audiobook playlists.
As these systems become more influential, readers will reasonably want to understand why certain books appear.
Is the recommendation there because it genuinely fits the request?
Because the book is popular?
Because the publisher has a commercial relationship with the platform?
Because the listener has behaved in a certain way previously?
Those questions already exist around recommendation algorithms.
Conversational AI may make them more important because a conversation can feel much more like receiving advice than scrolling through a list.
That makes transparency increasingly important.
The reader's relationship with the book is becoming less static
I have been thinking about this for some time in relation to AI Book Companion™.
The printed book is still enormously valuable.
The audiobook is valuable.
Neither has to disappear for something new to emerge around them.
What seems to be changing is the assumption that the reader simply consumes the author's thinking and then works out what to do with it alone.
Spotify's latest features provide another small piece of evidence that reading and listening are moving towards something more interactive.
Readers can describe what they want.
AI can help them discover the right book.
Conversation can continue while they are listening.
The next question is who should govern that conversation when the reader wants to go deeper into the author's ideas.
For an expert author, I think the answer increasingly points back towards the author's own intellectual property.
That is where the difference between general AI and a genuine AI Book Companion™ becomes much more important.
Sources
Spotify — Plot Twist: 3 New Ways to Discover and Explore Audiobooks on Spotify, 11 August 2026.
Alliance of Independent Authors — coverage of Spotify's AI audiobook discovery rollout, 15 August 2026.
I use AI to help research, organise and refine my writing. The ideas, expertise and intellectual property are my own, and every article is reviewed and approved by me before publication.