🐴 An overview of AI in FigJam

Walkthroughs and impressions

May 11, 2024

Happy Saturday!

Today, from sunny New York, I’m going to talk about AI in FigJam.

I use FigJam almost every day as my go-to online whiteboard tool for sharing ideas and brainstorming with others.

It’s hard to believe that it has only been about half a year since Figma launched its AI features in FigJam. It feels much longer than that…

If you have not used those AI features, then this article is a comprehensive introduction for you to get onboard. If you have already used those features, this can provide a deeper dive.

“…people sometimes freeze when faced with a new file—what we’ve been calling the “blank canvas problem.” We wanted to change that…”

- Noah Levin, Vice President of Product Design, Figma

I resonated with what Noah said. The first step is always tedious and time-consuming.

Before those AI features, I had to either do the initial setup by manually adding sticky notes or look for the right template online and grab what I needed. Not to mention the amount of time I spent moving things around just to make the whiteboard looks good.

There features of AI in FigJam can be put into two buckets:

Generate
Sort & Summarize

Let’s dive in!

Generate

“Generate” focuses on creating things from scratch, such as meeting templates and diagrams, in only a few clicks.

The items it can generate include:

🧠 Mind map
⏳ Timeline
➡️ Flow chart
📊 Gantt chart
📈 Org chart
👥 1:1
📝 Plan
💭 Brainstorm
🗓️ Weekly sync
🔄 Team retro
🔄 Process flow
🌳 Family tree

I’ll choose three featured ones for today’s demo.

1) Mind Map 😀

Our team is going to have an onsite week, so the first thing that came to mind was creating a mind map of onsite team activities.

I was impressed by the mind map it created for me in just a few seconds. It not only provided more ideas than I expected but also categorized them into logical groups.

2) Flow Chart 🙂

I asked ChatGPT to generate a comprehensive “if yes, then; if not, then” decision tree, which I then used as a prompt to feed into FigJam. I was happy with the result too—a nice first step that made it easy for me to refine based on it. The color-coded visuals was really nice.

3) Brainstorming 🙂

The result it generated for the vision workshop was impressive. I always love how visually organized the AI-generated results are. It would have been a tedious task to manually make small adjustments just to make the boards look neat. The photo widget in the Icebreaker was very fun to play with if you haven’t tried it before.

Then I added one line to the previous prompt, “A voting session at the end”, hoping to make revisions to what FigJam AI had generated for me.

Surprisingly, it created a totally new set of boards next to the previous set. The outcome appears to be a random new generation.

An area of improvement I noticed is that it is almost impossible to iterate or fine-tune based on the generated outcome.

If I wasn’t happy with anything, I had to manually edit it or keep generating new ones—that started to become time-consuming as well.

Sort & Summarize

“Sort & Summarize” focuses on taking the information from sticky notes and then sorting or summarizing it accordingly.

1) Sort 🙂

I was too lazy to create all the sticky notes from scratch in FigJam for the demo, so I did the following:

First, I asked ChatGPT for a prompt like “Give me 20 pieces of user feedback about a meal planning app”.
Then, I asked ChatGPT to convert the bullet points into a table.
After that, I copied the table and pasted it into FigJam, ensuring it was also in table format there.
Next, I searched for the “Table to Sticky Notes” plugin and ran it in FigJam.
Finally, I selected the table and clicked on “Create”.

Here we go! Each piece of user feedback is now on a separate sticky note.

Next, I selected all the sticky notes and clicked on “Sort stickers”.

It ended up categorizing all the sticky notes into three buckets, and the titles were accurate.

Note that the sorting result is often not as accurate as this example. The user feedback in this case was generated by AI. In the real world, however, the sorting results often need careful review and manual adjustments.

It is good enough to have the initial sorting, but there are often outliers that the AI has trouble understanding correctly.

2) Summarize 😔

This is what I received when I selected all the sticky notes and used the “Summary” feature.

While the summary makes sense as a whole, I was disappointed that I couldn’t change any aspect of this summary—it was a static image.

It might be sufficient for a quick and scrappy note during a brainstorming session, but it's far from something official.

Takeaways

There are many features under the "Generate" bucket. I’ve been actively using less than 10% of them. The rest, such as goals, timeline, weekly sync, and Gantt charts, I never got to use. I can see product managers using them more often than designers. And for many of these features, since I usually have a clear format in mind, I didn’t need to bother using FigJam AI to help me kick-start.
I don’t trust the "Sort" and "Summarize" features in FigJam yet. They can be used as the first step, but I had to scrutinize the results and move things around. Chances are, I can’t use them directly without extensive editing. I don’t think AI is to blame here. Information on sticky notes is raw by nature, and we often write convoluted information on them. These are all reasons that make it even harder for AI to sort and summarize information effectively.
Overall, FigJam’s AI features are centered around generating templates and categorizing information. If you want to perform tasks related to "creating ideas", then these features are not a good match. I recommend checking out my previous article on Jambot in that case.

Thanks for reading. If you enjoyed this issue, please consider sharing Design with AI newsletter to someone who might benefit from it.

Have a great weekend! I am going to my kid’s birthday party :)

—Xinran

P.S. Amazon released my 3rd book The Portfolio Secrets the other day. If you know a designer who is working on portfolios, that will be a helpful book with in-depth analysis and examples.