Interrogatory LLM

Martin Fowler describes a technique called 'Interrogatory LLM' where instead of manually writing context for an LLM task, you prompt the LLM to interview you with one question at a time to gather the necessary information. The resulting context document can then be used in a separate LLM session. A second use case involves giving an LLM a specification document and having it interview a domain expert to verify accuracy — useful because people often find reviewing documents harder than having a conversation. Fowler credits Harper Reed's blog for the original description and notes the technique is broadly applicable for extracting knowledge from people who struggle with writing.