Proposition · Agent taxonomy
Types of AI Agents
Agent Types hub — bento map and journey strip. Classic grid-only layout remains at /en/agent-types/.
From reactive to orchestrated
- 1 Reflex
- 2 State
- 3 Goals
- 4 Utility
- 5 Learning
- 6 Multi-agent
Bento map
Six cells, one mental model — pick by complexity and outcome.
Simple Reflex Agents
Rule-based agents that react only to current input, without memory.
Open guide
Model-Based Reflex Agents
Reflex agents with internal state to track partially observable environments.
Open guide
Goal-Based Agents
Agents that evaluate actions based on whether they move toward a defined goal.
Open guide
Utility-Based Agents
Agents that score outcomes and pick the action with the best utility trade-off.
Open guide
Learning Agents
Agents that improve policies over time from feedback and experience.
Open guide
Multi-Agent Systems
Multiple specialized agents coordinated to solve complex tasks.
Open guide
Classic layout
Prefer the earlier grid-only hub? It stays available at the same URL as before.