We set out to answer a simple but deep question: Can AI actually practically help product designers improve during the discovery and ideation phase of the design process?

So we spent 5 weeks running an experiment.

We mapped every tool we use for discovery: Mobbin, Dribbble, Pinterest, Twitter, Behance We broke down typical design thinking and brainstorming workflows We reviewed every prototyping or idea-capturing tool we’ve used Then we tried building lightweight AI workflows with various LLM tools and frameworks

Result: Yes. Used well, AI can significantly improve design thinking — especially for junior/mid-level designers — by offering faster idea generation, design critiques, and creative merges.

Out of that research, we built Moonchild: A discovery-stage design ideation tool that:

Generates thoughtful UI concepts from minimal prompts Allows asking design questions and getting structured critique Merges styles, flows, and interaction patterns from multiple directions Outputs great Figma-ready screens and UX flows, fast Try it (private beta): https://moonchild.ai Use code 'hackernews' for early access.

Would love feedback — especially from product designers, PMs, and UX folks doing early-stage work.