An expert system is software that attempts to mimic human expertise by searching through a database of known answers or solutions in order to help solve a problem. Chains of decision points known as decision trees are analyzed by human experts and coded into the system to help the software identify the correct area of solutions. How the expertise is embedded into the knowledge and the branching decisions is key to determining the success of the system.
Since a domain of expert knowledge can involve a large number of facts and decision branches, most expert systems are confined to concentrated areas to minimize the size and scope involved. Most experts systems don’t involve the ability to learn, but it is possible to incorporate some ability to expand either the knowledge base or the decision branches, or to join an expert system with another type of AI software in a hybrid system.