The COMPEL Glossary Graph visualizes relationships between framework terminology, showing how concepts interconnect across domains, stages, and pillars. Term nodes cluster by pillar affiliation while cross-references reveal semantic dependencies — for example, how risk appetite connects to control effectiveness, model governance, and assurance requirements. This network representation helps practitioners navigate the framework vocabulary and understand that COMPEL terminology forms a coherent conceptual system rather than isolated definitions.
COMPEL Glossary / online-evaluation
Online evaluation
Assessment of an AI system under live traffic using randomized or sequential experimental designs — A/B test, multi-armed bandit, canary, or interleaving.
What this means in practice
The only evaluation mode that measures true user-facing outcome; governance constraints include blast radius, reversibility, and regulatory exposure during the live test.
Synonyms
online test , live-traffic evaluation , production evaluation
See also
- Offline evaluation — Assessment of an AI system against static datasets — training hold-out, validation set, benchmark corpus — without exposure to live user traffic.
- Multi-armed bandit — An online experimentation strategy that shifts traffic toward better-performing variants during the experiment — trading statistical power for exploitation of early wins.