Learn it. Teach it. Lead the change.
The Delegation Ladder is becoming a shared method for AI-centric change management. If you lead change, advise organizations, teach, or own AI in your company, this path gives you the vocabulary, the instruments and a credential earned on real work.
The practitioner program is forming now. Join the list to hear first.
A field with no shared method, and a lot riding on it
One language
One name and one scale for clients, practitioners and students: six rungs, zero to five. No crosswalks between competing scales.
Per decision, not per agent
Delegation is a contract for a consequential decision, not an agent-wide maturity score. That's what makes it teachable and testable.
Built for change
The six-stage journey doubles as a change plan: see the shift, decide who decides, build who can work, enforce, count, keep humans answerable.
Climb by doing
Each tier builds on the one before. You move up by producing real artifacts on real work, the same way autonomy is earned on the ladder itself.
Everyone
Foundations
The vocabulary, the six rungs, and the five plain questions every AI use should answer. One hour to one day.
You'll be able toPlace a decision on a rung and say who put it there.
Reviewers
Reviewer
What to check, with scheduled time and the standing to say no. Practice on real register entries.
EvidenceReviews completed on live register entries.
Owners and sponsors
Owner and sponsor
Writing grants and Intent Contracts, running the ratchet down, and reading both ledgers.
EvidenceA signed grant and a ledger readout.
Builders
Builder
The Consequence Layer, the Rule-to-Control Matrix, the Prove It test deck and the decision-scoped grant schema.
EvidenceMatrix rows that passed their tests.
Teach the teachers
Facilitator
Licensed to teach the method inside a firm or a university, using the facilitator kits.
EvidenceA cohort you have led through the journey.
Credential the evidence, not the knowledge
You can't certify judgment with a multiple-choice test. So the credential is awarded for artifacts that ran on real work, and it stays consistent with the ladder's own rule: autonomy is earned, not configured.
- Earned on real work. A signed grant, Matrix rows that passed tests, a ledger readout.
- Renewed on a cadence. Methods and models change; the credential keeps up.
- Revocable. Like any grant on the ladder, it can be decided again.
Channels already in place
Be first on the trail
Tell us which tier interests you and where you'd apply the method. We'll reach out from info@bytebrain.org when the founding cohort opens.
