Product Design in the age of Figma Make

In a world where AI is everywhere, I’ve been experimenting with how tools like Figma Make can fit into my design workflow. I’ve found myself wondering: is this just a gimmick, or can it actually help me work faster and create better experiences for our users?

The blank canvas is intimidating. As a designer, I’ve struggled with that first step of a new project. Lately, I’ve been exploring how Figma Make can help. By simply turning an idea into a prompt, I can get a rough prototype in seconds. This isn’t meant to replace a thorough discovery process — sometimes you still need a FigJam to untangle data, user needs, and business goals — but it’s an incredible tool for bringing an idea to life. Now, I can explore a concept in minutes rather than hours.

Sometimes the biggest hurdle in a project isn’t the time it takes to create a design itself but getting stakeholders to see the vision. With a quick prototype, I can bring an idea to life in a way a static wireframe never could. It turns a concept into a conversation starter, making it easier to gather feedback and build momentum. Suddenly, you’re not just describing a solution — you’re showing one. This is key when you’re trying to get buy-in and tell a story.

My process

I start with a prompt in ChatGPT. (I’ve tried asking ChatGPT to write a prompt designed for Figma Make, although I haven’t noticed any difference in the quality of output this way. Let me know if you have!) In this prompt, I give AI a high-level problem or a user need, and it generates a prototype for me. I’ve learned that trying to get the prototype to be perfect at the start isn’t necessary. Initially, I tried feeding it our design system library, but the results were almost always worse; the output felt stiff and forced. I’ve found that the real value of these tools isn’t in generating a polished, final design, but in giving me a real-feeling site I can edit.

Here’s a breakdown of my process:

1. Start with a prompt: I give AI a problem that I’ve defined through research and examining business needs. Once I have my prompt, I add it to Figma Make. From there, Make generates a working prototype I can use as a starting point. Here’s an example of a prompt:

Design a desktop and mobile web interface for the athenaHealth Partner Listings page in athenaConnect, where Partners can manage and publish their Marketplace app listings. >

2. Adjust and refine : I take the output and begin adjusting it, not so much to make it “pretty,” but to make it functional. I’m focusing on the user flow and the core interactions, not pixel-perfection or using design system components. I try to remove distracting components or flows that are unnecessary to the key personas for which I’m designing.

Figma Make tip: Use the “Point and edit” tool to select specific elements of the UI to edit. This allows you to quickly update things like text color & size, or to give Figma Make a more specific reference of what you want it to edit.

The “point and edit” feature is the curson icon.

Once selected on an element, you’ll see a toolbar where you can make changes manually or switch to the AI prompt tool.

3. Get it in front of users, fast: Once I have a working prototype, I put it in front of as many users as I can. This is where I learn the most. I want to see if the core idea works before I invest much time refining it. To do this, I take the prototype and post the link in community forums for our users. I also run user interviews with a handful of participants, or use AI- moderated tests to expedite research (more on that soon)!

4. Use it as a blueprint: If the concept resonates with users, I then hand it off to engineering. The deliverable? A clear blueprint of the flow and the intended user experience. From there, they can build it using our existing, high-quality design library components versus trying to work through AI’s messy code or components. Alternatively, they can work with other software to convert the Figma Make code into workable code with our design system components.

Some considerations

Until Figma Make is able to replicate specific design systems, this process works best at companies where pixel-perfect visual design isn’t the top priority. The outputs from Figma Make are a fantastic blueprint, but they’re not a final product. There has to be an understanding that developers will need to swap out the AI-generated components with those from real design systems. There’s also still a need for a design QA to ensure the vision is translated properly.

Another thing to consider are AI credits. I’ve had to incorporate spend into the design process when using other AI vibe coding tools and a credit limit is coming sometime later this year for Make (according to Figma). If you’re using this tool, make sure your design leadership is aware of the upcoming changes.

Final thoughts

In the end, this process isn’t about replacing my design skills with AI. It’s about freeing myself from the blank canvas problem and tedious prototyping so I can focus on what truly matters: understanding the user, shaping the journey, and making sure the final product is both usable and valuable.

So, how are you using AI to free up your own creative process?