The most important terms around AI agents – explained clearly for decision-makers.
AI Agent
An autonomous AI system that pursues goals, makes decisions and takes actions – without constant human input. Unlike chatbots, agents act independently.
AppFunctions
Technology that lets apps expose their functions directly to AI agents. This allows an agent to operate an app without "seeing" the screen.
Autonomous Execution
An AI agent's ability to complete multi-step tasks on its own – even when the user is not actively involved. A core property of modern agent frameworks.
Agentic AI
AI systems that don't just answer questions but actively get things done: searching the web, writing code, sending emails, calling APIs – fully autonomously.
RAG (Retrieval Augmented Generation)
A technique where AI agents first retrieve relevant information and then use it for more precise answers. Combines web research with language models.
Durable Execution
An agent's ability to run long-running tasks reliably – even through connection drops or restarts. A core feature of modern agent frameworks.
Multi-Agent System
Several specialized AI agents work together to achieve complex goals. Each agent has a clear role (research, analysis, communication, execution).
Sandbox Computing
AI agents run in isolated environments to minimize security risks. Prevents an error in the agent from affecting the entire system.
Context Window
The maximum amount of information an AI model can process at once. Modern models offer up to 1 million tokens of context – enough to grasp entire codebases or document archives at once.