liveBook · Manning

Model agency

As explained in Agent Design Patterns

Model agency is the form of agency in which AI-model outputs guide the selection of the next action.

Model agency occurs when AI-model outputs guide an agent’s next action. In a hybrid workflow, control can pass between model decisions and code that executes predefined steps—a pattern called code agency. section 1.1.1

Personal assistant

A workflow can pass from code invoking speech recognition, to a model interpreting a request, to code accessing a calendar, and back to a model generating a contextual reply. section 1.1.1

How Agent Design Patterns covers it

How deep it goes
Defines model agency, contrasts it with code agency, and traces a personal-assistant handoff without detailing implementation.
Examples
Uses a personal assistant that moves from speech recognition to request interpretation, calendar access, and a contextual reply.
What you can do after
Understand how decision-making can pass between model outputs and executing code in an agent workflow.
How it connects
Connects agency to AI models, code agency, agents, and orchestration.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Agent Design Patterns. The section links point to the parts of the book it draws on.