Case-based reasoning

Case-based reasoning (CBR) solves novel problems by retrieving and adapting proven solutions from a library of similar past experiences.

CBR operates on a four-step cycle (Retrieve, Reuse, Revise, and Retain) to bypass first-principles logic in favor of historical precedent. Pioneered by Roger Schank at Yale in the early 1980s, the paradigm transitioned from cognitive theory to practical AI with Janet Kolodner’s CYRUS system. It powers diagnostic tools: medical platforms match patient symptoms against thousands of clinical records, while legal databases use it to surface relevant case law. By indexing specific episodes rather than rigid rules, CBR excels in domains where data is incomplete or highly context-dependent (like technical support or real-time navigation).

https://en.wikipedia.org/wiki/Case-based_reasoning