Agency
Agency is the capacity of an entity to make choices and initiate or carry out actions with some degree of autonomy, often in pursuit of goals.
AI Agents in Action, Second Edition
Agency is the capacity to perceive, decide, and act with some degree of autonomy in pursuit of a goal.
- How it is used
- Distinguishes a goal-directed, multistep agent from an assistant that completes a request and stops.
- How deep it goes
- Defines agency operationally and develops its decision loop, tool use, and safety boundaries without pursuing philosophical criteria for intention.
- Examples
- Maps image creation, travel booking, and computer buying to tasks and tools, then contrasts a prompt-only research planner with tool use.
- What you can do after
- Evaluate/understand it by identifying when multistep tool decisions warrant autonomy and where validation, budgets, or approval should constrain it.
Read how AI Agents in Action, Second Edition explains agency →
Introduction to Generative AI, Second Edition
An agentic capability is the ability of an AI system to plan and carry out actions toward goals with limited direct human intervention.
- How it is used
- Frames AI as systems that perceive, plan, and act across workflows rather than merely generate predictions or responses.
- How deep it goes
- Explains agency through capabilities, safeguards, and consequences, while deferring implementation details and precise autonomy thresholds.
- Examples
- Uses workflow agents, semiautonomous collectives, hybrid task systems, autonomous drones, and smart appliances as illustrative cases.
- What you can do after
- Evaluate agentic systems by considering their capabilities, failure modes, oversight requirements, and effects on responsibility.
Read how Introduction to Generative AI, Second Edition explains agency →
Delivering Machine Learning Projects: From design to deployment
Agency is the capacity of a person or system to make and carry out decisions that affect an unfolding situation.
- How it is used
- Protects meaningful human intervention when systems delegate control over unfolding, potentially harmful situations.
- How deep it goes
- Defines agency through intervention timing and authority, while deferring detailed requirements for autonomous-system development.
- Examples
- Uses life-support control, aircraft stall intervention, and a remote-control rover control-flow diagram.
- What you can do after
- Evaluate whether a human handoff occurs early enough and grants enough authority to prevent harm.
Read how Delivering Machine Learning Projects: From design to deployment explains agency →
Related concepts
- intelligencecontrasts with · Intelligence is the capacity to acquire and apply knowledge, reason, learn, solve problems, and adapt to changing situations.
- AI agentpart of · An AI agent is a computational system that uses artificial intelligence to interpret information, select actions, and pursue goals through interaction with an environment.
- neural networkcontrasts with · A neural network is a computational model of interconnected, parameterized units that learns transformations from inputs to outputs.
- toolproduced by · A tool is an instrument or software capability that enables a person or system to perform a task, obtain information, or interact with resources.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.