Spotify is staking out a strong position on AI-assisted development.
During its fourth-quarter earnings call, co-CEO Gustav Söderström said the company’s top developers “have not written a single line of code since December.” Implementation work is now handled by generative AI systems, with engineers guiding, reviewing, and approving the output.
The company tied this approach to measurable product activity.
Spotify shipped more than 50 features and updates across 2025, including improvements to music discovery and new tools that add context within the app, along with expanded audiobook functionality.
Recent launches include AI-powered Prompted Playlists, Page Match for audiobooks, and About This Song.
Supporting this workflow is an internal platform called “Honk,” which integrates Claude Code into everyday engineering processes.
Why It Matters: Spotify is describing a setup where generative AI tools are built directly into the systems that create and release its software. At the same time, the company is shipping more features and developing its own music data while setting policies for AI-generated songs.
- AI as the Primary Implementation Engine: According to Söderström, the company’s strongest developers are no longer typing out production code. Engineers describe the task in natural language and rely on AI to generate the changes. Their role now focuses on guiding the system, reviewing what it produces, and determining when the result is ready to ship. The process turns coding into oversight of AI-generated work.
- “Honk” Connects AI to Deployment Infrastructure: Spotify’s internal system enables engineers to interact with Claude Code directly through Slack. Söderström described a scenario in which an engineer on a morning commute messages Claude from a phone to fix a bug or add functionality to the iOS app. Once the AI completes the task, a new build is delivered back through Slack for inspection and approval. The engineer can merge the update into production before arriving at the office. This ties conversational prompts to live deployment pipelines.
- Sustained Feature Output Across the App: More than 50 features and changes were released in 2025 across major areas of the app. Updates touched music discovery and audiobooks, along with new contextual tools inside the listening experience. Prompted Playlists lets users create playlists from text prompts, while Page Match links audiobook suggestions to a listener’s preferences. About This Song adds background details directly within the app. Spotify credited AI-assisted coding and deployment for helping increase the pace of these releases.
- A Proprietary Dataset Built on Taste and Context: Executives highlighted Spotify’s music dataset as a competitive asset. Unlike factual information that can be scraped or summarized from public sources, music preference data is shaped by culture and personal context. Workout playlists in the United States often lean toward hip-hop, while EDM may be more common in parts of Europe and heavy metal among many Scandinavian listeners. Spotify said this type of behavioral data exists in large quantities on its platform and improves each time internal models are retrained.
- Policies for AI-Generated Music: Analysts asked how the company is handling AI-created tracks. Spotify said artists and labels can indicate in metadata how a song was made, including whether AI tools were used. At the same time, the company continues to monitor uploads for spam and abuse. The approach allows AI-assisted creativity while maintaining oversight of content quality and platform integrity.
Trusted insights for technology leaders
Our readers are CIOs, CTOs, and senior IT executives who rely on The National CIO Review for smart, curated takes on the trends shaping the enterprise, from GenAI to cybersecurity and beyond.
Subscribe to our 4x a week newsletter to keep up with the insights that matter.
