Top product managers view designing with AI as a holistic process. Instead of focusing solely on prompt engineering, they consider the entire workflow: understanding constraints, leveraging different AI tools for specific tasks, and maintaining human oversight to ensure quality and empathy.
Instead of facing a blank canvas, create a custom GPT that asks a series of structured questions (e.g., product goal, target user, key flows). This process extracts the necessary context to generate a focused, high-quality initial prompt for prototyping tools.
For design exploration, Google's Stitch tool offers a "YOLO mode" that pushes the AI to generate wild, unconventional design options based on an initial concept or screenshot. This is a powerful technique for breaking out of incremental improvements and exploring truly novel solutions.
A powerful, free workflow combines two Google tools. Use Stitch for divergent, visual ideation by generating multiple design variations from a prompt or screenshot. Then, export the preferred design directly to Google AI Studio to instantly convert it into an interactive, code-based prototype.
AI can generate designs but fundamentally lacks human empathy. This creates risks of bias and generic solutions. "Designing consciously" requires keeping humans in the loop to validate insights, double-check sources, and ensure the final product truly serves user needs.
Separate your workflow into two steps. Use a less expensive model like ChatGPT for the conversational, clarification-heavy task of building the perfect prompt. Then, use the more powerful (and costly) Claude model specifically for the code-generation task to maximize its value and save tokens.
When generating an initial prototype with AI, explicitly instruct the model to ignore standard features like sign-up or login. This forces the AI to concentrate its efforts on the key user flow that directly solves the user's core problem, leading to a more valuable first iteration.
Before using a dedicated AI prototyping tool, run your prompt through Claude.ai first. Its artifact generation provides a quick, lightweight visual of the prompt's output, allowing you to catch errors and refine the prompt without wasting time or credits on a more robust platform.