Glossary of AI Agent & Model Terms

Agent Architecture is the internal structural framework that defines how an intelligent agent perceives information, processes data, makes decisions, and performs actions within an environment. In artificial intelligence (AI), software engineering, and multi-agent systems, agent architecture represents the conceptual and technical design that determines how...

Agent evaluation refers to the systematic assessment of how well an AI agent performs in real-world or simulated environments. Because AI agents combine probabilistic reasoning, tool usage, and dynamic decision-making, their behaviour cannot be fully validated through static testing alone. Evaluation frameworks are therefore...

Short definition  Agent Memory is the mechanism that allows an AI agent to store, retrieve, and utilise information across time, including both short-term task context and long-term persistent knowledge.  Extended definition  Agent memory refers to the systems and processes that enable an...

Agent Orchestration is the process of coordinating, managing, and controlling multiple intelligent agents so they can work together effectively to complete complex tasks or achieve shared objectives. In artificial intelligence and multi-agent systems, orchestration refers to the structured organisation of agent interactions, workflows, decision...

AI Agent — a software system built on large language models (LLMs) that can reason over time, use tools, maintain state, and execute multi-step tasks toward a defined goal. Unlike a standalone LLM that generates a single response, an AI agent actively plans actions, interacts...

Autonomous Agent is an AI agent capable of independently planning, executing, and adjusting actions over time without continuous human input, while operating within predefined constraints, policies, and objectives.  An autonomous agent is a category of AI agent designed to operate with a high degree of independence...

LLM-powered Agent is an AI agent whose core reasoning and decision-making capabilities are driven by a large language model (LLM), combined with tools, memory, and control logic to execute multi-step tasks toward a defined goal.  An LLM-powered agent is a type...

MCP (Model Context Protocol) is an open standard that helps AI applications connect to external tools and information. It gives an application a consistent way to discover what a connected service can provide and request it when needed. Imagine asking...

Multi-Agent System (MAS) is a distributed computational framework composed of multiple autonomous software agents that interact within a shared environment to achieve individual or collective goals. Instead of relying on a single central intelligence, a multi-agent system distributes decision-making across independent...

RAG (Retrieval-Augmented Generation) is a way to help an AI model answer questions using information it can look up, rather than relying only on what it already knows. Think of an employee asking an AI assistant, “What’s our current refund...

Short definition  Tool-Using Agent is an AI agent capable of invoking external tools, systems, or services to extend its capabilities beyond internal model reasoning and execute real-world actions.  Extended definition  A tool-using agent is a type of AI agent designed...