Spotify

Justin Reock, Deputy CTO at DX, lists Spotify among the companies that pointed AI at the right bottleneck: it accelerates incident resolution by pushing context automatically to its SRE teams. Reock builds that argument on the theory of constraints (Goldratt), where an hour saved on a task that is not the bottleneck buys nothing, and Spotify sits beside Morgan Stanley, which reverse-engineers legacy Cobol, and Zapier, whose onboarding bots cut ramp-up from 90 days to two weeks.

Bill Staples, GitLab's directeur général, reaches for the same company toward a different end. In his essay of 24 August 2026, the experiences of Stripe, Spotify and Amplitude are the raw material for showing where the next constraints reappear once generation stops being scarce: environment, CI, review, governance. A thirty-minute pipeline, he writes, defeats any model.

The Wardley mapping guide places Spotify at the far right of the evolution axis. Music runs from the first MP3s in genesis, to the first iPod as custom-built, to smartphone music apps as product, to Spotify and Apple Music as commodity: expected to be there, noticed only when they break. Spotify is also the creator of Honk.

So the same name carries two incompatible readings. Reock counts the incident-resolution gain as a bottleneck correctly identified; Staples counts that class of experiment as evidence that the queue has merely moved downstream, into CI, review and governance.