The use of AI to write code is leading to productivity gains which would’ve felt like science fiction even a year ago.
During Spotify’s latest earnings call, co-CEO Gustav Söderström revealed that the streaming giant has developed an internal tool called “Honk” that enables engineers to ship features to production while commuting on buses and trains. The system allows developers to request bug fixes or new features for Spotify’s iOS codebase using Claude, Anthropic’s AI assistant, directly through Slack on their mobile devices.
The workflow is remarkably streamlined: An engineer messages Claude through Slack to describe the feature or fix they need. Claude generates the code changes, which are then pushed back to the engineer as a QR code. The engineer can scan the code to test the new feature on their phone immediately. If satisfied, they can merge the changes to production—all without opening a laptop.
“This is speeding us up tremendously,” Söderström said during the call. He described the tool as part of a broader effort to “retool the entire company for this age” of AI development.
According to Söderström, Spotify engineers have already shipped approximately 50 features using this mobile-first, AI-assisted development process during their morning commutes. The company has been told by AI partners that this internal workflow represents “industry-leading” work in AI-assisted software development.
The system reflects a fundamental shift in how software can be built. Traditional development workflows require developers to be at their desks with full development environments set up. Honk collapses that process into a conversational interface accessible from anywhere.
Söderström emphasized that Spotify views this not as the end of AI-driven development innovation but as “just the beginning.” The company is investing heavily in the technology and expects to not only maintain but accelerate its pace of feature shipping.
“We’ve been embracing and investing in this technology evolution for some time, and it’s allowing us to move with much higher speed,” he said, framing the transformation as capturing opportunity amid change.
The revelation offers a glimpse into how AI coding assistants are being deployed inside major tech companies—not just as productivity aids for individual developers, but as integrated systems that fundamentally reshape development workflows and timelines. While GitHub Copilot, Cursor, Claude Code and similar tools have become common in engineering organizations, Spotify’s Honk system appears to push the boundaries of what’s possible when AI coding assistance is deeply integrated into company infrastructure and communication tools. And for an industry increasingly focused on shipping velocity and developer productivity, Spotify’s experience suggests that AI-powered development tools may deliver on their promise of transformational efficiency gains—even if engineers are standing on a crowded subway car.