Where documented work concentrates, and where automation would pay for itself. Based on 34 published processes (illustrative) — every dollar figure is an estimate range driven by the assumptions below, and ranges narrow as more processes are published. Recalculated whenever a process is added or updated.
Every figure below reacts to these. Defaults are a starting point for discussion, not findings.
Number provenance: observed measured from recordings (step durations, systems touched) · assumed upload metadata or the sliders above · modeled arithmetic on the other two — inherits their uncertainty
Solid bar = conservative end of the range; faded extension = aggressive end. Payback at the current build-cost assumption.
Pick a candidate, set what automating it would cost — see when it pays for itself.
Candidate steps by stage — from flagged in documentation to automated in production
Annual savings (midpoint) vs implementation effort assumed — top-left is "do these first". Hover for detail.
Systems touched across published processes — where tool spend meets real usage
Published processes per team — where documentation is deep, and where it's thin
Processes where one person is the only executor on every step (from RACI) — continuity risk no savings number captures
The pilot's own before/after
Attach license spend to each system and see cost per documented use — answers "where is our tool spend going" directly. Needs license data; not in pilot scope.