Knowledge-based systems (KBS) use human knowledge to solve problems normally requiring human intelligence. To succeed they should possess performance at the same level as the human expert. These systems have been used in industry, finance and government for many years. Escalating workloads, cost constraints and evidence-based medicine create the need for such tools in the clinical laboratory. A number of systems exist in chemical pathology and haematology. Approaches have varied from simple rule-based systems to more complex models using fuzzy logic and artificial neural networks and incorporating probability theory, pattern recognition and multi-variate analysis techniques. In the hands of enthusiasts these systems have been judged to operate satisfactorily (produce a correct differential diagnosis) in 85–90% of patients. The KBS shell is a software environment containing a knowledge acquisition system, the knowledge base itself, inference engine, explanation subsystem and user interface. The core components are the knowledge base (human knowledge represented by e.g. If-Then rules) and the inference engine (forward or backward chaining). The explanation subsystem, which renders the rules transparent, is an important attribute for user acceptance. System development is an iterative process through trial and error. Quite apart from a perceived threat to future employment of ‘experts’, these systems are not in more widespread use because of i) failure to examine user requirements adequately, ii) inadequate sources of acceptable organised knowledge, iii) a multiplicity of systems which cannot communicate with each other, and iv) a lack of agreed standards for coding and information exchange. The additional value of these systems as standardising, training and educational tools should not be overlooked.