Adaptive prompting is a technique in which a model modifies or expands its prompts as task information changes. Instead of relying on one fixed instruction, the system asks clarifying questions, refines its goals, and creates subquestions during a multistep interaction. section 7.6
Travel planning
A travel-planning agent may begin with “Where would you like to go?” After a vague answer such as “somewhere warm,” it can ask about budget, activities, and travel constraints before producing tailored results. section 7.6
The technique makes prompting an evolving interaction rather than a one-time input. It helps the model construct relevant context, although the description remains conceptual and does not specify implementation details or guarantees of reliability. section 7.6
