Measuring maturity in AI

Team,

I’m writing this as a note to our future selves, and as an invitation to be deliberate about how we get there.

For a long time, I’ve described design as a spectrum: art on one end (craft, emotion, interaction nuance) and science on the other (systems, flows, behavioural mechanics, frameworks). That spectrum still holds, but the centre is collapsing. Not because design is narrowing, but because it’s expanding downwards in the two extremes at once.

On the craft end, in the industry designers are gaining real technical ownership of the front-end experience: prototyping with live components and real data, contributing interactions and polish closer to production, and using AI tools to reduce UX debt instead of creating more output. The final 5% where quality lives stops being a backlog item and becomes something we can shape directly.

On the systems end, the work stretches further into strategy: mapping coherent narratives across features, making sense of complex journeys and hierarchies, aligning UX outcomes to business goals, and applying behavioural thinking to build systems that guide users consistently. If one part of the team goes deeper into execution, another must zoom out to ensure the whole experience hangs together.

AI, and especially agentic development, has accelerated this shift. We’re no longer limited to describing an experience and hoping it survives translation. We can prototype behaviour that’s close to reality, test it credibly, and make the work concrete early. But the future isn’t evenly distributed. Some teams are using AI to compound capability through real workflows, real components, better decisions, and better outcomes. Others are using it as a way to produce more without improving the system of work underneath. The gap is widening fast, and “we’re using AI” is becoming as meaningful as saying “we use the internet.”

That’s why we need a maturity model. Not as a new career ladder, not as new expectations, but as a shared language for us all. A way to look at reality and navigate the change together.

Because the opportunity in front of us isn’t strategy versus execution. It’s strategy and execution, held together as a team.
Strategy without execution becomes commentary. Execution without strategy becomes noise. We need the connective tissue, principles, narratives, systems and we need the ability to make things real enough that quality is palpable, felt, measured, and shippable. Not by turning every designer into an engineer, but by investing in tooling, enablement, and expectations so high-fidelity, system-aligned work becomes normal rather than exceptional.

This maturity model is how we start that shift on purpose.

Grab the model on Figma Community

The model moves from Limited, where AI is mostly personal and low-risk, to Reactive, where teams dabble with vibe-coded prototypes that look promising but don’t hold up. In Developing, usage becomes intentional and repeatable, with shared patterns and prototypes that improve alignment and decisions. Embedded brings system-aware, production-shaped work, tighter cross-functional collaboration, and more attention to trust, uncertainty, and boundaries. Leading is where capability compounds: clear tool choices, strong technical fluency, shared rules and feedback loops, and AI experience quality treated as part of everyday product craft.

That’s the future I want us building toward, and the path I want us to walk together.
This is the first of a few open letters where we’ll talk about evolving our process and the education and enablement tracks we’ll run to get there.

Generatively,
Matt