Agent Taxonomy
Types of AI Agents
Practical taxonomy of the most common AI agent types, from classical reflex agents to multi-agent systems.
Need
Simple automation
Use simple reflex agents for deterministic trigger/action rules.
Need
Multi-step planning
Use goal-based agents when tasks require path-to-goal decisions.
Need
Team orchestration
Use multi-agent systems for role specialization and parallel work.
Goal-Based Agents
Agents that evaluate actions based on whether they move toward a defined goal.
Read type details →
Learning Agents
Agents that improve policies over time from feedback and experience.
Read type details →
Model-Based Reflex Agents
Reflex agents with internal state to track partially observable environments.
Read type details →
Multi-Agent Systems
Multiple specialized agents coordinated to solve complex tasks.
Read type details →
Simple Reflex Agents
Rule-based agents that react only to current input, without memory.
Read type details →
Utility-Based Agents
Agents that score outcomes and pick the action with the best utility trade-off.
Read type details →