Some of the best ideas come from unexpected connections.
When Sparkbox’s design systems and AI expedition teams discovered their work overlapping, they found new opportunities to learn and experiment together.
Bringing different perspectives into the same conversation helps us challenge assumptions, expand our expertise, and make room for discoveries we didn’t plan.
Some of the best connections are the ones you don’t plan.
At Sparkbox, our expedition teams give us space to experiment with new approaches and deepen our skills outside of client projects. We started with separate teams focused on design systems and AI, each exploring different challenges and opportunities within our work.
I have the opportunity to lead our design systems expedition team, where our shared goal is to deepen our design system knowledge across Sparkbox. We wanted to establish a common baseline for designers and developers while creating space to experiment with emerging tools and approaches in both design and code.
This kind of collaboration is especially valuable in design systems work. A decision that makes sense in Figma might introduce complexity in development. A technical constraint might reveal an opportunity to rethink how a component is designed. Having both designers and developers in the conversation helps us see the bigger picture and learn from perspectives we might otherwise miss.
Building a design system for sparkbox.com gave us a practical environment to put those ideas into practice. We worked through audits, foundations, components, and documentation while exploring how AI could support different parts of the process. From experimenting with Claude skills to speed up audits to finding ways AI might help bridge Figma and code, new opportunities kept emerging.
Meanwhile, our AI expedition team was exploring LLMs, building agent skills, and looking for practical ways to apply emerging technology to everyday work. They were testing tools and exploring how AI could make familiar processes more efficient. Many of those experiments have already found their way onto the Foundry.
As our conversations evolved, something interesting started happening.
The topics kept overlapping.
Two separate areas of exploration were starting to ask the same questions.
The value came from what we could discover together.
Some of the best connections are the ones you don't plan.
When Two Conversations Become One
Over time, it became more and more clear that there was more value in exploring them together.
That’s when we decided to merge the two expedition teams into one.
The interesting part wasn’t simply that our topics overlapped. It was what happened when people with different expertise started looking at the same problems.
Familiar challenges suddenly had new possibilities.
Take design system documentation, for example. Maintaining accurate documentation across design and code can be a time-consuming process. Our design systems team was thinking through the challenges of keeping everything organized and up to date. Meanwhile, the AI team was exploring how LLMs could work with existing information and make it more useful.
Bringing those conversations together led us to explore how AI could audit design system artifacts across Figma and code, identify missing or inconsistent information, and help turn that information into documentation that’s easier to understand and maintain.
Another opportunity was maintaining parity between design and code. How might AI help identify inconsistencies across different parts of a design system before they become larger problems?
These weren’t necessarily solutions we had figured out. They were questions we could explore together, building on the knowledge and experiments each team had already started.
And that’s where the value of combining the teams became clear.
People with more experience experimenting with AI could recognize opportunities to improve familiar design system processes. At the same time, those working more closely with design systems could bring practical experience and a deeper understanding of the challenges involved.
As a designer, I’ve found that some of my most valuable learning happens when developers challenge how I’m thinking about a problem. They bring questions and considerations I might not have recognized on my own. Adding people with different AI experience to those conversations opened up even more ways to think about the work.
The benefit wasn’t simply learning about another tool or gaining more technical knowledge. It was developing a better understanding of how our different areas of expertise could work together.
Neither group needed to have all the answers.
Creating the Conditions for Serendipity
We didn’t start our expedition teams with the intention of combining them. Each group had its own direction, and the overlap emerged naturally as we explored our respective topics.
What started as two separate conversations became an opportunity to learn something neither team might have discovered independently.
Serendipity isn’t something you can schedule. You can’t predict exactly when different ideas will connect or what those connections will lead to.
But you can create the conditions that make them more likely.
For us, it started with making room for exploration outside of client work. Expedition teams gave us dedicated space to experiment with ideas we were curious about, without needing to have everything figured out from the beginning.
It also meant bringing different perspectives into the same conversations. We had designers and developers exploring design systems, people interested in AI, and people approaching similar challenges through different areas of expertise. Those shared interests made it easier for conversations to cross over naturally.
Having something real to work on made a difference, too. Building a design system for sparkbox.com gave us practical challenges to explore. Rather than discussing ideas only in theory, we could experiment with actual components, documentation, and workflows.
And perhaps most importantly, we were willing to follow where those conversations led. When we recognized that the two teams were exploring similar questions, we didn’t need to keep them separate just because that was how we’d started.
There’s a lot of expertise spread across our team. Some of it comes from client work, some from individual interests, and some from the experiments we pursue simply because we want to learn something new.
The opportunity lies in the way that knowledge comes together.
Sometimes, you just need to get those people and ideas into the same room and see what happens.
