Design systems caught me at the right time. I’ve long felt I straddle the worlds of design and code, and with a little experience, moving to a more product-oriented space, enabling bringing these worlds together and more. This stuff isn’t easy and often isn’t what it appears to be. A well-formed library in Figma is no more a design system than a collection of components in React. The unseen parts, the bits around how people work, function, and relate are where the bulk of the effort tends to go.

Their perceived simplicity can lead us into a false sense of security; this is a done deal and we just need to get to a point like [insert well-known company]’s design system and we’ll’ve smashed it. That’s not the case. All of the parts within a design system are contextual to your environment (structure, leadership, size, skillsets, experience, goals, etc) as well as moving and evolving, as our understanding of things like design tokens or how to express design changes and adapts. We as a community push each other to keep improving and adapting to challenges, which is so awesome to be a part of…but I have my concerns.

Out of all the design systems that people work on, how many would be said are successful at the moment? Anecdotally, those that attempted this many years ago might be on their 3rd or 4th attempt. Typically, we’re doing the right thing in systemising design to generate better outcomes for websites, and yet this is still difficult. There are few examples of multi-brand or multi-platform design systems working well (which is a monster challenge) and yet that’s the space they can be most impactful. Dealing not only with responsive design (don’t get me starting on how dysfunctional design-developer relations on the use of device-based breakpoints are) but multiple platforms as we aim to express our brands not just on the web but into native apps as well…how will we cope with and adapt to more future change?

All of this work is based upon doing what we already do, mostly making websites, but better and more efficiently. Looking at how we might use these same processes if not the artifacts and outcomes for related platforms or features feels a long way off. Done well, I believe the hub of a design system isn’t design tokens or components but a collective, working together, making decisions together, and applying brand through various means to solve user and business problems. This means that we can and should not only bake in the capacity to change and adapt within this space but the systems we create as a result of it. This enablement and empowerment of product or feature teams can be a powerful example of symbiosis when done right.

“AI” is on the hype cycle at the moment, so it’s as good an example as any to use. What we mean by AI at the moment is a wide range of tools, most of which aren’t actually AI…certainly they don’t have intelligence. This might be machine learning to language models. Wildly different inputs and outputs. If our organisation opts to embrace something like this — what might that mean?

Part of what we can do, I think, is recognise that our artifacts are transient. Today it might be components and tokens, tomorrow it may be something else. That should challenge the notion of what we mean by a design system. The best quotes that sum up (in a reductionist way) our current state of defining them revolve around it being design, code, people, and the relationships between them. If we move into spaces that don’t require design tokens in the same way or using different design tools or code, that kind of definition still holds true.

I’m a believer that in various ways, some of our future tools and outputs might be multi-modal. For example, today ChatGPT is a form that accepts text and gives you a text response. It’s a preview of the technology, not the end goal. If you could type, talk or show an image to your tool and flip between modes of communication, they may become more of an extension of us than a thing we go to for requesting an output. The phrase ‘co-pilot’ that some are using already feels like a pretty good fit in this regard. Bake in modes of interaction with the intelligence AI can give us and that’ll be a pretty amazing space.

Stepping back to consider the service design that underpins many of our larger apps or websites, and bringing in the possibility of some of these kinds of future changes for our end users, what customer touch-points might use one or more modes? Today, there’s only a little talk about using design systems for emails, where there’s a huge opportunity there to expand our design system work to another customer-facing medium. Being less constrained to how we think of a design system means we can welcome not only emails but whatever comes next and not need to start over.

Looking at future changes means in some ways being comfortable with the unknown and giving us space to adapt and evolve. We shouldn’t need a new design system for a new platform or mode of communication necessarily (though there can be times that can make sense), just appreciate that what we consider to be a design system today might not be what they become tomorrow. Allowing for us to experiment and challenge that base idea, I think might give us the scope to meet whatever challenges come our way…after all, they’re just design, code, people, and the relationships between them all.