Aadil Ghani

We keep talking about AI agent costs as: model price tokens requests

But agentic systems add another cost: uncontrolled execution.

• An agent retries a tool 5 times.

• Calls an expensive model when a cheaper one would work.

• Loops through unnecessary steps. Hits APIs it didn't need.

• Delegates work to another agent.

I m a UI/UX designer and recently started exploring how AI can help bridge the gap between a Figma design and a more functional prototype.

Traditionally, I d create the UI in Figma and then build interactions separately. With the newer AI tools, I m curious how designers are changing this workflow.

For those who work with AI-assisted prototyping:

• How do you go from Figma designs to functional prototypes?

• Which parts of the process does AI handle well?

• Where do you still prefer traditional Figma prototyping?

• Have you found any tools that preserve the original design particularly well?

Well, after this post, they probably won t be that secret anymore.

But there are still a few things not everyone knows about: