Building scalable mobile applications through AI-driven development

Few companies have gone as far with AI-driven engineering as Spotify. The company has shared publicly that almost all of its engineers now code with AI, that pull request frequency has risen by around 60%, and some of its most experienced developers haven't written a line of code by hand since late 2025. Much of that runs through Honk, Spotify's internal background coding agent, which can pick up a task, write the change and verify it in CI across platforms.

For our consultant, a senior Android engineer, that has changed what a working day looks like. Less time goes into typing Kotlin. More goes into framing problems precisely, breaking work into tasks an agent can own, reviewing what comes back and making the calls no model can: architecture, user experience and quality for an app used by hundreds of millions of people.

It's a shift our consultant has seen before, from the other side: as an early evangelist for Jetpack Compose, Google's declarative UI toolkit for Android, helping several teams get up and running with it. Moving teams from one way of building to another, and making the new way feel safe and practical, turns out to be exactly the skill AI adoption demands.

There’s a deeper link too. Spotify has pointed out that agents perform best in codebases that are consistent and follow shared patterns. Declarative UI with clear, repeatable conventions is precisely that kind of code.