Agent Definition translates the AI’s role into an explicit behavioral contract. It defines what the AI is responsible for, how it should interpret evidence and apply business logic, what decisions or recommendations it is permitted to make, and what it must produce in response.

The outcome is a shared definition of expected AI behavior that can be translated into scenarios, requirements, acceptance criteria, edge cases, and evaluation tests.

What's Included:

• Responsibilities

• Operating Logic

• Decision Rights

• Outputs

Agent Definition turns product intent into sufficiently explicit AI behavior for detailed requirements to begin. The team defines the AI’s responsibilities, the operating logic that guides how it performs them, the decision rights that establish what it may infer, recommend, initiate, or defer, and the structure and requirements of its outputs. Operating logic captures the business definitions, rules, criteria, evidence standards, and exception conditions the capability needs to interpret its work consistently within its intended boundaries. Together, these definitions provide the foundation for decomposing AI behavior into scenarios, business rules, user stories, acceptance criteria, edge cases, and evaluation tests.