Some context: Quotient allows developers to evaluate, improve, and ship high-quality AI products with fast, real-world, data-backed experimentation. In the quickly changing (and saturated) AI product market, they needed to provide an exceptional developer experience to convert and retain customers—engineers building and refining AI systems for production use.
Designing an elegant prompt refinement loop
PromptLab was designed to be clear and intuitive. A clean, structured layout guides users through prompt refinement and output evaluation. Key elements (model parameters, variables, prompts, and results) are organized for quick scans and low cognitive load.
A neutral palette and subtle highlights help direct attention without distraction. Suggestions from IQ (Quotient’s AI copilot) are presented in a way that feels intuitive and actionable.
The flow is built around simplicity and iteration. Users define prompts and variables, run the model, and provide human feedback on outputs. Then, based on user feedback, IQ generates targeted suggestions—from wording tweaks to structural changes—helping refine the prompt for improved results.
Crafting seamless micro-interactions
With many micro-tasks happening on a single screen, we relied on micro-interactions to make the experience fluid and intuitive, while making sure user confidence remained high as they navigated the sometimes complex prompt-building flows.
How? Subtle animations, instant feedback, and quick hints mean users can navigate complex workflows without leaving the main view or second-guessing their actions:
