An intelligent agent is a sub-system of the artificial intelligence but with a lot less functionalities and "intelligence". An intelligent agent is simple a software tool designed to assist end users in performing non-repetitive task related to computing processes. The intelligent part is that it can act on its behalf when configured to respond to specific events. An intelligent agent is sometimes referred to as bot (short for robot) and is often used for processing assistance in data mining.

There are two general types of intelligent agents – the physical agent and the temporal agent. The physical agent refers to an agent which uses sensors and other less abstract and more tangible means to do its job. On the other hand, the temporal agent may be purely codes that use time based stored information which are triggered depending on configuration.

From those two general types of intelligent agents, there may be five classes of intelligent agents based on the degree of their functionalities and capabilities. These five are simple reflex agents, model-based reflex agents, goal-based agents, utility-based agents and learning agents.

The simple reflex agent functions on the basis of its most current perception is also based on the condition – action rule such as "if condition then action rule". The success of the agents job depends on how fully observable the environment is.

The model based agent has its current state stored inside the agent which maintains certain structures describing the part of the world which are unseen and this kind of agent can handle environments which are partially observable. The behavior of this kind of agent requires information about the way that the world works and behaves and thus is sometimes considered to have the world view model.

The goal based agent is actually a model based agent but it stores information about certain situations and circumstances in a more desirable way by allowing the agent some good choices from among many possibilities.

The utility based agent uses a function that can map a state to a certain measure of the utility of the state.

Finally, a learning agent are is a self governing intelligent agents that can learn and adapt to constantly changing situations It can quickly learn even from large amounts of data and its learning can be online and in real time.

Intelligent agents have become the new paradigm for software development. The concept behind intelligent agents have been hailed as "the next significant breakthrough in software development". Today, intelligent agents are used in an increasingly wide variety of applications intended for a wide variety of industries. These applications range from comparatively small systems such as email filters to large, open, complex, mission critical systems such as air traffic control.

In large data intensive applications like internet web servers, an example of an intelligent agent would be a software for determining ranks of websites. This is very important because ranking is the basis for advertisement rate and the overall value of the website. An intelligent agent may be used to audit the websites ranking in the leading search engines around the world.

Perhaps the most ubiquitous example of an intelligent agent is found on our computers. These are in our anti virus and anti spyware systems. Intelligent agents are constantly on the lookout for viral strain updates so our computers can be protected all the time.