Not an AI problem.

The situation

A marketing team asked us to point AI at their case-study problem. Case studies weren't getting written, the library was thin, and AI looked like the obvious fix.

What we found

When we mapped how the work happened, the blockage sat upstream of anything a model could touch. Nothing triggered a case study when a piece of work finished, so nothing arrived for AI to draft. Automating an empty pipeline produces automated emptiness.

What changed

We took the AI build off the list and fixed the pipe instead: a simple capture step at the end of every significant piece of work, owned by a named person, triggering the case study, the PR angle and the client feedback request in one step. Once material started flowing, AI drafting became the easy part, and this time it had something real to work with.

Point AI at a working process and it flies. Point it at a missing one and it automates the gap.