The capture-vs-design gap
Why do Scribe libraries
quietly rot?
You rolled out Scribe (or Tango, or Guidde) across the team. Everyone was supposed to record their work. Six months later the library is enormous, half the recordings are out of date, three different people documented the same task in three different ways, and new hires still end up asking a human rather than hunting the guide.
The library didn't fail because the tool is bad. It failed because the operation underneath it was never designed. These tools record what happens; they do not tell you which tasks deserve a recording, which roles own which work, which steps should have been automated, and which ones shouldn't exist at all. Without that design layer, you end up with a beautifully indexed museum of how your team currently copes.
Henry is the design layer. He maps your operation using the AAAERRR methodology from Deliberate Work, specifies each step across three planes, and runs a Fit Assessment. Only steps that belong in the Work Zone — work humans should continue doing, at least for now — get recorded with Scribe, Tango, or Guidde. The rest get automated, delegated to AI, or retired.
For operations teams and fractional COOs building systems that survive, this is the difference between a documentation library and a designed operation that happens to include guides.
Record only the Work Zone
Henry's Fit Assessment flags which steps belong to humans. Those are the ones worth recording. Everything else shouldn't be a Scribe in the first place.
Guide titles from step intent
Work Plane intents become Scribe titles and opening summaries. Readers know what they're about to do, not just which buttons to click.
Ownership from role specs
Execution Plane roles become Scribe collection permissions and owners. Each recorded guide has a human accountable for keeping it current.