Behavioral Alignment · Shape
Behavioral Alignment · ShapeAutonomy Scale
Interaction-control protocols across the levels of autonomy: from simple adjustments to advanced customisation of the system's decisions.
Why it matters
Users who can modify system behaviour as needed without losing trust: agency designed across both data and AI-output control, not left as an afterthought.
The canvas
The worksheet you fill in, as a template. Use it with your team here, or fill it in online in the workspace.
Project Team Date
12345678910
Full human agencyFull system agency
What your lowest level means
Level 1 (Full Human Agency): What must the user do entirely by hand?
What your highest level means
Level 10 (Full System Agency): What does the system execute silently, with zero notification?
Default action
Name the specific task (e.g., Tagging incoming support tickets)
Level (1–10)
Why this level?
Why this level? What is the user's default expectation, and what is the blast radius if the AI misinterprets the input?
Low-stakes action
Name the specific task (e.g., Drafting suggested meeting agendas)
Level (1–10)
Why this level?
Why this level? How easily can an error here be dismissed, ignored, or undone in one click?
High-stakes action
Name the specific task (e.g., Permanently deprecating user permissions / executing a payout)
Level (1–10)
Why this level?
Why this level? If the system hallucinates or fails here, what is the cost (financial, legal, reputational, or data loss), and why is automated execution unsafe?
app.jcerejo.com · The Anticipatory Design Framework
Fill it in online: start freeFree: one framework project. No card.