Glossary
Autonomy Spectrum
The range from human-in-the-loop assistance to fully autonomous AI action.
Definition
The autonomy spectrum is the framework describing how much independent action an AI system is permitted, from suggestion-only assistance, through human-approved actions, to fully autonomous execution with human oversight by exception. Where a workflow sits on the spectrum is a deliberate design and governance choice.
Context
Mapping each workflow onto the autonomy spectrum is one of the first practical decisions in any agentic-AI deployment. High-stakes, regulated, or reputation-sensitive work stays toward the human-in-the-loop end; high-volume, low-risk work earns more autonomy over time as evidence accumulates. The right placement balances efficiency against control.