An AI agent is software that uses AI to interact with the environment. Extremely versatile, AI agents can not only access external sources such as a database or website, but they can learn by trial and error. The key to any agent, AI or otherwise, is that it is continuously monitoring a situation and can act when certain events occur. An AI agent can even begin processing based on the time of day. Unlike generative AI, which requires a prompt from the user, an AI agent takes action on its own. See
agent and
reinforcement learning.
AI agents are becoming popular because they add human-like decision making to traditional computer-based tasks. They not only handle routine functions such as handling customer requests or booking a trip but they can automatically refine their algorithms to improve their effectiveness the next time.
Agent Terminology
One or more AI agents can be part of an AI system that includes development, training and generating answers. An AI agent is also called "agentic AI" and an AI system is also called an "AI model" or just plain "AI." See
system and
AI model.
Replacing Everything?
In late 2024, Microsoft CEO Satya Nadella said AI agents are going to be part of all software applications in the future. Considering that Microsoft has made a fortune in business applications, Nadella's comment sparked a lot of controversy. Nevertheless, by all accounts, we seem to be heading into a very different technological future. Stay tuned! See
AI anxiety.