ORBiS TREE · ENGLISH DOCUMENT
Concepts →AI Agent
An AI system that selects and executes actions over multiple steps to pursue a goal.
한국어
English
This English page is a curated translation layer linked to the Korean source article. Community changes are currently made on the Korean source, where the full revision history and anonymous edit trail are preserved.
An AI agent is a system that uses an AI model to decide what to do, take actions, observe results, and continue toward a goal. The term covers many designs, from simple tool-calling loops to systems with planning, memory, and multiple specialized components.
How it works
A typical agent loop combines a model with instructions, available tools, state, and an execution environment. The system may search, call APIs, manipulate files, or interact with software before deciding on the next step.
Why it matters
Agents can automate longer workflows than single-turn chat, but autonomy increases the cost of mistakes. Permission boundaries, sandboxing, monitoring, and recovery become important when an agent can affect external systems.
Related concepts
SOURCES
Sources
-
ReAct: Synergizing Reasoning and Acting in Language ModelsPrinceton / Google Research / arXivOpen source ↗
KNOWLEDGE LINKS