Eight weeks, three phases, every DC CAP team.
Every hour Claude saves on drafts and data pulls goes back to the roughly 800 scholars we serve. Ten senior leaders across six units proved the model from April to June 2026; now every team learns it, funded by a $600K grant from the KPMG AI Impact Initiative.
These are the goals the organization holds itself to. Nothing on this list is graded or turned in by any one person; the course stays self-paced and open.
Most staff reach confident, regular use of Claude on real work, visible in the aggregate of the assessments staff choose to take.
Confidence grows across the course, and the staff who take to it early begin teaching the rest.
Each unit produces at least one champion candidate.
Units collect before-and-after examples from real work; each unit redesigns two workflows.
At least one unit ships a capability that could not exist without AI.
Zero Tier 1 or Tier 2 data incidents; every output has a human owner.
Full goal detail and measurement framework: KPI Framework (built for the spring 2026 leadership pilot) | Governance goals: AI Governance Framework
Every team has put Claude on real work at least once
Every unit can name at least one workflow where Claude saves time or improves quality
Fluency growth visible in the aggregate of assessments staff choose to take; 3+ reusable workflow templates documented; each unit ships one working build
