Dynamic Prompt Middleware: Contextual Prompt Refinement Controls for Comprehension Tasks
Ian Drosos, Jack Williams, Advait Sarkar, Nicholas Wilson
TL;DR
This work tackles the difficulty users face when prompting GenAI for comprehension tasks (e.g., explaining formulas or code) and introduces dynamic prompt middleware to provide context-sensitive controls. It presents two prompt control approaches—Dynamic Prompt Refinement Control (Dynamic PRC) and Static Prompt Refinement Control (Static PRC)—implemented as an Option Module that generates refinements and a Chat Module that grounds responses with those refinements, arranged in a two-tier architecture with inline and session options. A formative study with $n=38$ participants identifies a trade-off between standardized, predictable prompting and adaptive but potentially opaque controls, guiding the design goals of controllable and adaptable interfaces. A controlled within-subjects study with $n=16$ shows a clear preference for Dynamic PRC, which increases perceived control and task exploration but raises questions about cognitive load and interpretation of option effects; the work concludes with design implications for future Dynamic PRC systems and discusses alignment with broader explainability goals and the potential for Generative UI to augment human reasoning in knowledge work.
Abstract
Effective prompting of generative AI is challenging for many users, particularly in expressing context for comprehension tasks such as explaining spreadsheet formulas, Python code, and text passages. Prompt middleware aims to address this barrier by assisting in prompt construction, but barriers remain for users in expressing adequate control so that they can receive AI-responses that match their preferences. We conduct a formative survey (n=38) investigating user needs for control over AI-generated explanations in comprehension tasks, which uncovers a trade-off between standardized but predictable support for prompting, and adaptive but unpredictable support tailored to the user and task. To explore this trade-off, we implement two prompt middleware approaches: Dynamic Prompt Refinement Control (Dynamic PRC) and Static Prompt Refinement Control (Static PRC). The Dynamic PRC approach generates context-specific UI elements that provide prompt refinements based on the user's prompt and user needs from the AI, while the Static PRC approach offers a preset list of generally applicable refinements. We evaluate these two approaches with a controlled user study (n=16) to assess the impact of these approaches on user control of AI responses for crafting better explanations. Results show a preference for the Dynamic PRC approach as it afforded more control, lowered barriers to providing context, and encouraged exploration and reflection of the tasks, but that reasoning about the effects of different generated controls on the final output remains challenging. Drawing on participant feedback, we discuss design implications for future Dynamic PRC systems that enhance user control of AI responses. Our findings suggest that dynamic prompt middleware can improve the user experience of generative AI workflows by affording greater control and guide users to a better AI response.