Concepts
What is an AI Agent?
The term "agent" appears across computer science, law, economics, and philosophy—each discipline bringing different assumptions. This framework provides a precise, cross-disciplinary definition through six operational properties organized into a three-level hierarchy.
The Three-Level Hierarchy
Not all agents are created equal. This hierarchy distinguishes basic agents from production-ready systems and AI-powered implementations.
Foundational
Agent
The minimal definition: any system with Goal, Perception, and Action capabilities. A thermostat qualifies.
Production-Ready
Agentic System
Adds Iteration, Adaptation, and Termination for multi-step tasks with learning and graceful stopping.
AI-Powered
AI Agent
An agentic system whose capabilities are powered by AI/ML, particularly large language models (LLMs).
The Six Properties
Agency is defined by six operational properties, divided into foundational (GPA) and operational (IAT) categories.
Goal
The objective or purpose that guides behavior. Can be explicit instructions, implicit preferences, or emergent from training.
Perception
The ability to observe and interpret the environment through sensors, APIs, or data sources.
Action
The ability to effect change through actuators, API calls, or tool use. Actions can be reversible or irreversible.
Iteration
Execute multiple perceive-act cycles, building on prior state and environmental feedback.
Adaptation
Modify behavior based on experience, feedback, or changing conditions—within or across sessions.
Termination
Recognize when to stop—goal completion, resource limits, errors, or need for human escalation.
Why This Framework Matters
For Governance
Different levels require different oversight. A Level 1 thermostat needs minimal governance; a Level 3 AI agent handling legal research needs robust controls, escalation paths, and audit trails.
For Communication
Provides shared vocabulary across disciplines. When a lawyer and an engineer discuss "agents," this framework ensures they're talking about the same thing.
For Evaluation
Before deploying any "AI agent," ask: Does it have all six properties? If not, which are missing? The answer reveals what risks remain unaddressed.
For Design
Building an agent? The six properties become your checklist. Missing Termination? Your agent may run forever. Missing Adaptation? It can't learn from mistakes.
Related Glossary Terms
Agent
A system exhibiting the three foundational properties of Goal, Perception, and Action (GPA). An agent pursues objectives, observes its environment, and takes actions to achieve its goals. This represents Level 1 in the three-level hierarchy.
Agentic System
A system exhibiting all six operational properties: Goal, Perception, Action, Iteration, Adaptation, and Termination (GPA+IAT). Agentic systems are production-ready and can operate across multiple cycles with learning and graceful stopping. This represents Level 2 in the three-level hierarchy.
AI Agent
An agentic system (Level 2) whose capabilities are powered by artificial intelligence or machine learning, particularly large language models (LLMs). This represents Level 3 in the three-level hierarchy.
GPA (Goal, Perception, Action)
The three foundational properties that define minimal agency. Goal provides direction, Perception enables environmental awareness, and Action allows the system to effect change. Together, they form the basis for all agentic behavior.
IAT (Iteration, Adaptation, Termination)
The three operational properties that distinguish production-ready agentic systems from basic agents. Iteration enables multi-step execution, Adaptation allows learning from experience, and Termination ensures graceful stopping.
Three-Level Hierarchy
The conceptual framework distinguishing three levels of agency: Level 1 (Agent) with GPA properties, Level 2 (Agentic System) with all six GPA+IAT properties, and Level 3 (AI Agent) where capabilities are AI-powered.
Learn More in the Book
This framework is developed in depth in Chapter 1 of Agentic AI in Law and Finance, including historical context, cross-disciplinary perspectives, and practical applications.