article --summary
The summary
This episode of the Marketing Ships to Production conversation follows Tanya Higgins as she describes turning her brand assets into an AI-assisted hub that grew into an internal tool at Tiger Data. Her experience shows how a designer on a small team can move from waiting on engineering for every asset to building and deploying things herself.
Stephan Schmidt frames the new bottleneck as deployment review, where AI speeds up building but everyone still waits in a queue to reach production. He warns that the same speed creates shadow apps and unmanaged deployments, and he argues that the answer is to review permissions rather than every line of code, much like how phone apps request access to the camera. Matty Stratton adds that good guardrails should assume people will make mistakes, and that platform engineering becomes more valuable in smaller organizations.
The conversation ends with advice for engineers building these systems: they should be patient with non-engineers who are nervous about production and make the right way the easy way.
Editorial note: This is a summary of an article first published by Arrested DevOps. The full piece is on their site.