🐴 Thoughts on Figma's new AI features
What surprised me and what didn't
☀️ Happy Tuesday!
“What do you think of Figma AI?” I received a message right after Config this year.
Figma’s new AI features have gotten a ton of attention from the design community right after the announcement and now it seems the heat has cooled down a bit.
I originally planned to talk about this in my July 5th newsletter, but I switched to discussing custom GPTs instead, since I had already talked a lot about AI in Figma.
Now, it’s time to get back to it.
Figma's AI features are currently in limited beta and will roll out gradually through the end of the year. If you're on a team or organization plan, your admin needs to opt in to enable these features, or you won't be able to access them.
Overall impression
What surprised me was the sheer magnitude of Figma's AI feature updates. I knew they were working on AI, but there ended up being more variety than expected. Each feature, though small, uses AI in unique ways to make designers' lives easier.
What didn’t surprise me, though, is that most of these AI features aren’t groundbreaking. Some features felt familiar, resembling third-party Figma AI plugins I’ve tested before. It looks like Figma is incorporating ideas from existing plugins into its own features.
The unsurprising ones
1️⃣ Rename Layer
Time to say goodbye to “Group 27” and “Layer 82 copy 5”…
Figma claims that this AI feature will contextually rename the layers with just a click. This helps simplify the traditional bulk renaming process.
🤔 This new feature reminded of Figma Autoname, a third-party plugin that I’ve test before.
It is a free tool for renaming layers. Better yet, it is open source. I could select many frames in Figma and the tool will automatically rename them.
For example, if the frame is a Card, then it will be renamed to “Card”. If the frame is a paragraph of text, then it will be renamed to “Paragraph”.
The renaming still feels generic though—not useful enough yet. Curious to see how effective the “Rename Layers” feature will be.
2️⃣ Replace text content with AI
It has two sub-features:
Rewrite this: It lets you generate text from scratch, or rewrite text.
Replace content: It uses text from the first element in a series of duplicated elements to populate content in the remaining elements.
🤔 “Rewrite this” reminded me of Write Better - AI Text Editor
It is one of the useful third-party text-rewriting plugins in Figma that I tested. And of course, you can do similar things in ChatGPT too if you don’t mind switching programs.
Regarding “Replace content”, I have not seen similar plugins in Figma. Curious to see how well it will perform.
3️⃣ Remove background
Figma offers this new feature to remove the background from any image. This comes in handy when you want to overlay an image over a design, such as when designing a product landing page.
🤔 This new feature reminded me of Remove Background (Privacy-First).
It is a simple, easy-to-use plugin with an impressive result. Worked smoothly for me without issues.
The surprising ones
4️⃣ Suggested auto layout
Figma claims this feature can analyze your design and automatically determine which objects should be placed in auto layout frames, creating them as needed.
I didn’t expect this feature. While Auto Layout is incredibly useful, it can get time-consuming as designs become more complex. I ran a quick test with a semi-complex layout, and within seconds, it generated the auto layout for me, creating three levels of hierarchy.
Overall, it looks promising. I haven’t tested it with more complex use cases yet, but I’m curious to see how well it performs. If it works as expected, it could save us a lot of time by creating optimized Auto Layout structures.
5️⃣ Generate designs from text prompts
The feature can transform ideas into editable designs with its built-in UI kits.
I was most excited about this feature in Figma and couldn’t wait to test it out.
While several other platforms have similar features, I wasn’t satisfied with them. Most of them required me to design and edit outside of Figma, which was inconvenient. For a couple of them that work within Figma, the generated designs weren’t as useful as expected at all.
However, after Andy Allen’s post on X accusing Figma of ripping off Apple’s Weather app, things changed quickly. Figma decided to pause the feature only several weeks after its announcement.
There was a long response from Dylan Field, Figma’s co-founder, on what happened. He claimed this feature was “powered by off-the-shelf models and a bespoke design system commissioned by Figma”.
6️⃣ Visual Search
This feature allows you to search for a related components or assets based on the image you provide.
In the short term, it will search through your team’s files, but Figma claims it will eventually be able to search across files from the Figma Community as well.
This feature can be especially valuable for larger teams, where there are more components to search through. While the upcoming ability to search Figma Community files is super exciting, it could also raise potential privacy concerns.
7️⃣ Quick-click prototyping
This feature claims to quickly turn static mocks into interactive prototypes.
The idea is interesting and something I didn’t expect at all, but it feels a bit too good to be true. I’m a little skeptical—if I end up having to manually make significant revisions after the auto-generation, it might not be as useful as it seems.
That’s it for today! What do you think of those features?
Let me know in the comment.
—Xinran
P.S. No one got the answer right from last week’s newsletter… The mystery item was a fresh walnut straight from the tree! Yes, it’s edible—and it tastes quite different from the ones you buy in stores.
🎉 Our 2nd cohort has launched!
Excited to announce that the 2nd cohort of AI for Product Designers has launched on Maven.
It is probably the most practical AI course you can find that is specifically tailored to product designers and UX/UI designers. You’ll discover tips and tricks to design smarter, engage in hands-on exercises, receive bonus resources, and connect with other like-minded designers.
The course is priced at just $399—more affordable than the other courses on Maven. Plus, the first 10 people to sign up can save $100 with the exclusive discount code EARLYBIRD100 at checkout. Secure your spot today before it’s gone!
(You can probably get your course reimbursed with your company’s L&D budgets. Use this reimbursement template for a smooth conversation with your manager.)
This post caught me up. I didn't go to Config or watch videos (because I’m overwhelmed). I look forward to try the text-to-UI and prototyping feature once they release.
I also think Dylan's response is very good in acknowledging his responsibility. The “I hate missing the mark” part makes it all very relatable.