Resources / RiGi Group
Glossary
A plain-language reference for the terms used across RiGi and WWAIOS.
Reference
Language for accountable AI
Decision authority
The person or policy permitted to approve or perform an action.
Evidence position
What supports a claim, and which relevant facts remain unknown.
Evaluation
A defined test of system behavior against stated criteria.
Guardrail
A check that limits or redirects unsafe or unauthorized behavior.
Lineage
A record connecting source, transformation, policy, and outcome.
Human handoff
A point where a person reviews or takes responsibility for the next step.
Retrieval augmented generation
A method of supplying relevant external sources to a model at request time.
Policy as code
Rules expressed in a form that can be versioned, tested, and enforced.