Expert Systems

Systems that represent domain knowledge explicitly, usually as facts and rules, and use an inference engine to reach conclusions. They were central to symbolic AI and remain appropriate when policy must be explainable, stable, and auditable.

An expert system does not automatically learn from every interaction and should not be confused with an LLM. Clear rules have maintenance cost and limited coverage; in return, they make it possible to explain which rule led to a decision. Hybrid architectures can use a model to interpret language and a rule system to decide or validate critical effects.