Human Agent Interaction

Defining how humans interact with agents

Purpose and Structure of this Guide

Comprehensive, practical, and expert-level resource based on the five core HAX principles.

Comprehensive, practical, and expert-level resource based on the five core HAX principles.

This guide focuses on five core HAX principles derived from these guidelines:

Let the Human Set the Rules

Show the agent's Reasoning

Assume the agent will make mistakes, make them clearly fixable

Support turn-taking, negotiation, co-editing, and human input

Let users audit, debug, and learn from the agent's behavior.