Promptions
Promptions (“Prompt” plus “options”) dynamically generates UI to help users steer AI responses more effectively. It’s simple, effective, and easily customizable, making it suitable for developers from individual vibe-coders to enterprise software engineers.
Try Promptions Repo
Read Promptions Paper
Promptions was released for research purposes. Users can learn, explore and experiment with Promptions.
For many users of chatbot-style AI systems, getting a good response is hard because writing all the context needed is both difficult and tedious. Promptions makes it much easier. Given a single simple user prompt, Promptions makes a behind-the-scenes AI call to generate parameterized options to refine the output, which are shown to the user as UI elements such as radio buttons, checkboxes, or toggles. As the user interacts with these elements, the AI response is re-generated in real time to show what effect different choices would have.
A Promptions chatbot interface (left) contains UI elements to steer the AI response in ways that would take much more effort with a traditional AI chatbot interface (right).
Promptions system model. (1) The Option Module ingests the user’s prompt input along with the conversation history. (2) It then outputs a set of prompt options, each initialized based on the content of the prompt. (3) These options are rendered inline via a dedicated rendering engine. (4) The Chat Module incorporates the refined options as grounding, alongside the original prompt and conversation history, to generate a chat response. (5) The user can modify the GUI controls, which updates the refinements and triggers the Chat Module to regenerate the current response accordingly.
Promptions is grounded in research that compares static UI elements for steering AI compared to dynamic elements. In the study, participants reported greater preferring dynamic elements, that it made them feel more successful with AI.
Comparison of participants’ preferences of dynamic choices (Promptions) to static choices.
Participants perceived the dynamic controls as more effective in managing AI output and expressed greater satisfaction with the degree of control provided, in contrast to the static condition, where many reported a continued need for additional control.
Comparison of participants’ reported effectiveness of dynamic choices (Promptions) to static choices.
Promptions is part of a growing movement to make AI more user-centered, customizable, and accessible. By giving users intuitive control over how AI responds, we’re building a future where everyone—from novice coders to seasoned engineers—can create smarter, more responsive AI experiences.
Promptions was invented by Ian Drosos (a past Microsoft Research Resident) with support from Jack Williams, Advait Sarkar, Nicholas Wilson, Payod Panda, and Sean Rintel. It is a collaboration between the Tools for Thought and ENCODE projects in the People-Centric AI focus area at Microsoft Research Cambridge, UK.