I attended AirOps Next in New York on May 13, 2026. Austin Lau from Anthropic discussed encoding workflows into shared skills. I kept coming back to the editorial side of that idea.
For a B2B content team using AI, what happens after an editor corrects a draft? Does the next draft benefit from that judgment, or does the editor have to make the same correction again?
Save the reason for the edit
Pick a workflow with a clear owner. When the reviewer changes an output, keep the reason with the correction. An unsupported claim needs a source check. A confusing explanation needs to be rewritten for the person reading it. The next run needs to know which problem the editor was fixing.
Then check the next output. Saving an instruction is only useful if the workflow applies it correctly. A mistaken instruction can spread the same error through every draft that follows.
This is the question I brought back from the session into my thinking about AI-assisted operations. I covered Austin's discussion in my May 14 recap.
The editor's reasoning should still be available when someone else takes over the work.