Platform teams that still ship
An internal platform dies when it becomes a ticket queue. Ours stay in the product calendar.
Most internal platforms fail the same way. A central team takes tickets. Product squads wait. The platform board fills with work that is technically correct and operationally useless.
We staff platform work as a product. There is a backlog that sits next to the customer backlog, a weekly review with the squads who consume the platform, and a rule: if nobody opened the new service this sprint, it is not done.
The artifacts are ordinary. A paved path for a new service. A golden path for observability. A written “how to get a production secret” that an engineer can follow at 11 p.m. without a Slack thread.