A conceptual framework introduced by Deep Decision Lab. It proposes that an agent's effectiveness can be explained through a common set of fundamental capabilities, providing a unified foundation for analyzing and improving decision-making across biological, organizational, and artificial agents.
CAT has a broad set of applications because it evaluates capability rather than domain knowledge. Here are several use cases:
Pre-implementation evaluation — Assess whether a proposed agent design has the capabilities required for success before investing in development.
Capability gap analysis — Identify which capabilities limit an agent's performance and prioritize improvements.
Agent benchmarking — Compare multiple AI agents, teams, or organizations using a common capability model.
System design validation — Evaluate whether a system architecture contains the necessary capabilities to achieve its intended objectives.
Failure diagnosis — Determine whether poor outcomes stem from deficiencies in perception, execution, governance, delegation, or other capability areas.
Human capability assessment — Analyze strengths and weaknesses of individuals for coaching, hiring, leadership development, or role alignment.
Organizational assessment — Evaluate whether an organization's decision-making capabilities support its strategic goals.
AI agent evaluation — Measure the readiness of autonomous AI agents before deployment into production environments.
Capability roadmap planning — Create an improvement roadmap by identifying the highest-impact capability investments.
Education and training — Design learning programs that target specific capability gaps instead of generic skill development.