Maturing a design system beyond components: operationally, culturally, and strategically
Everything I worked on last year to grow a design system, from variables and tokens to patterns and AI.
Now that I’m finally used to typing 2026 in date fields, it felt like the right moment to look back at 2025. In Brazil, we say the year only starts after Carnaval (I wish!), so I guess I’m not late for this kind of reflection 🤭
2025 was… something.
I started the year on full-energy mode, and ended it realising this would be my last full year focused on Design Systems (at least for now).
So here is an honest and emotional retrospective
.
My year at a glance
My focus throughout the year was on maturing our Design System beyond components and into something more operational, measurable, and sustainable. While still trying to catch up with all the AI buzz.
Some of the highlights (TL; DR)
Refined our style-based tokens library with variables and unified our two “core” component libraries into a single one (as it should be!), enabling theming in Figma for Product Designers. Such an exciting project!
Dived into Design System Patterns (one that
may have slightly traumatised meI learned a lot from), figuring out how to handle patterns more efficiently and collaboratively.Explored enablement and measurement, from tracking adoption and support requests to building internal AI tools like Nova, our DS bot assistant, making the system easier to use and understand.
A lot of my time also went into the “invisible” work: supporting product designers with complex UI decisions, audits, documentation clean-ups, and conversations that never turn into tickets but absolutely help shape product quality.
Learned (and also failed) different AI approaches along the way (a topic worth exploring another time).
Great feeling, looking back!
Now, if you feel brave enough, here are the not-so-short versions of my main work throughout the year.
Figma Design System Libraries 2.0
The work
Refining our previous style-based Design Tokens library in Figma by enabling Figma Variables, ensuring it matches code behaviour, and consolidating the two “core” libraries into a unified and themeable Core 2.0.
What I did
Refined the Design Tokens library structure by upgrading it to variables and aligning it with code behaviour, making it future-proof for upcoming projects.
Used that foundation to unify our previous two core libraries into a single, themeable component library in Figma.
Documented the migration with guides, a playground file, and video tutorials.
Bridged design and engineering through token refinements using Tokens Studio and PR reviews.
Outcomes
Product designers can finally work with themes properly in Figma, without hacks.
Tokens are more consistent, predictable, and aligned across design and code.
The system now has a solid foundation for multi-brand setups and platform expansion.
Philosophical take
Good design system work often looks invisible. If people can focus on design instead of figuring out library hacks, that’s success.
Design System Patterns
The work
Defining, documenting, and piloting Design System Patterns, turning an abstract idea into an actual practice that teams could use.
What I did
Defined what a pattern is (and isn’t) in the context of our design system.
Shaped the Phase 1 Patterns initiative, focusing on definition, documentation, and process.
Led the Contextual Feedback Indicators pattern from research to documentation.
Established documentation standards, a recommended (but flexible) pattern cycle, and improved overall DS documentation quality.
Outcomes
Teams can understand, discuss, and apply patterns.
Designers have clearer guidance on how to use and create patterns.
Design system topics moved closer to a shared responsibility, setting a good foundation for the decentralised model.
Philosophical take
Patterns are underestimated, and most of the work
should beis alignment and collaboration.
Enablement, Metrics, and Measuring Impact
The work
Making the system measurable and easier to understand.
What I did
Tracked adoption metrics for MDS 2.0 libraries, identifying gaps and opportunities.
Organised support request tracking from official Slack requests to identify frequency, track time spent, and identify recurring themes and opportunities.
Built and maintained Nova, our internal Design System AI assistant, connecting design system users to documentation and faster guidance.
Created additional DS-focused AI agent tools to support pattern definition and documentation work.
Outcomes
Designers have a better understanding and confidence in using the system.
Fewer repetitive support questions.
Clearer signals to prioritise improvements.
Templates and agents set a foundation for collaborative work.
Philosophical take
Metrics and AI don’t have the answer to everything, but they definitely help cut down on repetitive questions and figure out where to actually focus.
Collaboration and Support
The work
Supporting product designers and teams in real product scenarios, where the design system meets ambiguity, constraints, and deadlines.
What I did
Actively supported designers via Slack, reviews, and ad-hoc discussions.
Helped teams navigate complex UI decisions.
Improved and maintained documentation across Supernova and Confluence.
Participated regularly in design huddles and cross-team discussions to align on quality, accessibility, and system evolution.
Outcomes
Designers make better, faster decisions.
Higher trust in the design system as a partner, not a blocker.
Philosophical take
“Invisible” work scales. The small things might matter more than big, fancy-pants projects.
Closing the Year (and Opening a New Chapter)
(With a July update)
Toward the end of 2025, a company reorg led me into a new role as a Product Designer. Yes. Bye-bye, dear Design System team.
This honestly caught me by surprise, and meant saying bye to Design Ops as a team and some amazing design leaders.
Then, more recently, a new reorg made my new role redundant, too.
So instead of settling into "generalist Product Designer," I found myself with the opportunity to actually choose a path and bring the passion back.
And the direction is pretty clear to me now: exploring and improving the overlap of design systems in design, engineering, and AI.
If 2025 was an exciting year of hands-on work, 2026 is shaping up to be exciting in a whole new way.
New challenges
New direction
New user 🤖
Same systems-nerdy brain
Bring. It. On.