Work

Stretto · 2016–2026

Cadence Design System

A decade-long evolution from legacy modernization effort to enterprise design platform.

Serving five business lines and six teams (pods), Cadence became the catalyst for React modernization, Storybook adoption, design tokens, and the company's first dedicated UX Engineering practice, laying the foundation for AI-ready design system workflows.

Client
Stretto
Year
2016–2026
Role
Design system founder and manager, Principal Product Designer
Scope
Design Systems, UX Engineering, System Governance
Team
Product Design, Engineering, Product Management, Brand
Disciplines
Design Systems, Product Design
  • 5Business linesCadence grew from one product into a shared platform across five business lines.
  • Two dozenEngineers, roughlySupporting multiple product teams and roughly two dozen engineers.
  • 2First UX EngineersProposed in 2021; the company's first UX Engineers joined in early 2022.

What began as a visual design strategy for one product became the shared foundation connecting design, engineering, and product teams across five business lines.

Context

Modernizing more than software

In 2016, our team began transforming Best Case, a mature desktop application used by bankruptcy professionals, into a cloud-based SaaS platform.

The work appeared straightforward at first: modernize the experience, redesign aging workflows, and create a foundation for future growth.

As the platform evolved, it became clear we weren't just redesigning screens. We were building a shared system that needed to support products, teams, and technologies for years to come.

Legacy Best Case Desktop
Early Best Case Cloud wireframe

Growth created complexity

What began as a single-product modernization effort expanded significantly after Stretto’s acquisition and portfolio growth.

Cadence had to evolve from a design language for one product into a platform capable of supporting multiple business lines, brands, teams, and delivery models.

Timeline

CDS evolution, 2016–2026

  1. 2016

    Foundation

    Established the first shared visual language and UI standards while modernizing a legacy desktop platform for the web.

  2. 2019

    Stretto Acquisition

    Expanded the system vision to support a growing product portfolio and align multiple brands under one ecosystem.

  3. 2021

    Storybook Vision

    Validated component-driven development through Storybook and design-to-code pipeline proof-of-concepts.

  4. 2022

    UX Engineering Team

    Established dedicated UX Engineering support to scale adoption, governance, and system maturity.

  5. 2023

    Design-to-Code Pipeline

    Proved end-to-end implementation using React, Storybook, and a shared monorepo architecture.

  6. 2025

    AI-Ready Documentation

    Expanded documentation, content standards, and governance to support AI-assisted workflows.

  7. 2026

    MCP + LLM Integration

    Connected design system knowledge to AI agents through skills files, MCP, and structured documentation.

One product became a portfolio of products

In 2019, Stretto acquired CINgroup, expanding the product ecosystem and introducing a new challenge: Cadence could no longer serve only one application or one brand context.

The system needed to preserve what already worked while aligning with a broader portfolio of products, brand standards, stakeholders, and technical realities.

One business vertical became an ecosystem of verticals and products

Building through organizational change

The technical challenges were real, but the organizational challenges were just as, if not more, demanding.

Teams constantly reorganized, priorities shifted, and partnerships evolved across product, engineering, architecture, IT, design, and marketing. There was real apprehension and gatekeeping too, mostly from senior engineers who weren't familiar with modern framework and design system workflows.

None of that maturity happened in a straight line. Adoption depended on education first, which built trust, which eventually led to steady alignment rather than a mandate.

  1. CDS contribution model · © Zalando

    Accountability

    I would need to define ownership, decision-making, and contributions workflows across design and engineering teams.

  2. Release notes documentation

    Adoption

    I knew creating guidelines, release notes, office hours, and support channels would keep adoption visible and predictable.

  3. 2021 leadership alignment presentation

    Alignment

    Presenting during company-wide townhalls and All-hands meetings ensured I connected key stakeholders across all departments.

Approach

Designing a system not a library

We treated Cadence as a product, not a pile of components.

That meant envisioning and planning for scale, adoption, evolution, and transparency from the beginning.

  • Build for ScaleCreate foundations capable of supporting products, brands, and teams that don't exist yet.
  • Build for AdoptionMake the system easier to use than creating something new from scratch.
  • Build for EvolutionDesign for change, iteration, and growth rather than a fixed end state.
  • Build for TransparencyCreate visibility around ownership, governance, contribution, and decision-making.

Treating Cadence as a product

Cadence needed its own roadmap, backlog, release cadence, stakeholders, support channels, and success metrics.

The goal wasn't to make teams think about the system more. It was to make it dependable enough that they could think about it less and just fold it into how they worked.

New component reqs checklist for designers and developers

Governance before components

Years after the first standards were created, we formalized how work entered the system from designers, how decisions were made, who could/would/should contribute, and how accountability was shared across design and engineering.

Governance wasn't urgent in the early days, not until ambiguity got expensive and overcomplicated. I knew that once multiple projects and teams depended on Cadence, clear ownership would matter more than any individual component.

Architecting for shared connection

Cadence was never intended to be a static component library. The goal was a shared language connecting design, documentation, engineering, and product through a single source of truth.

Early architecture decisions were shaped by design system, atomic design, and content strategy thinking from the likes of Nathan Curtis, Brad Frost, and Jina Anne. Over time, those ideas evolved into a connected model for component ownership, token architecture, documentation, and implementation.

Original architecture and workflow model, informed by the work of Nathan Curtis, Brad Frost, and Jina Anne

Process

Documenting the foundations

As adoption expanded beyond Best Case Cloud, the original standards matured from scattered references in tools like WordPress, Axure, Notion, and Azure into a more formal foundation system.

That work turned visual standards into reusable tokens, themes, accessibility guidance, and implementation rules that could support multiple products, teams, and frameworks.

Establishing UX Engineering at Stretto

As Cadence expanded across products and teams, maintaining consistency required dedicated engineering capacity, not just components and documentation.

In 2021, I proposed and secured approval for the company's first dedicated UX Engineering function. By early 2022, two UX Engineers had joined the design organization to help scale adoption, accelerate implementation, and strengthen the connection between design and code.

This investment became a catalyst for broader frontend modernization efforts, including Storybook adoption, React migration initiatives, and the foundation of a design-to-code delivery pipeline.

Solidifying governance models

As more teams depended on Cadence, governance moved from informal alignment to operational practice.

We established a Jira backlog, RACI model, contribution workflows, and review processes so requests could be evaluated consistently across design and engineering.

The goal wasn't to add bureaucracy. It was to create a repeatable way to manage change without relying on institutional memory.

Operationalizing the system

A design system is only useful if teams can reliably find it, understand it, and use it.

To support adoption, we created documentation standards, release processes, support channels, office hours, roadmap visibility, and system health reporting.

These practices turned Cadence from a design initiative into an operating platform teams could actually rely on to deliver product work at scale.

Solution

Foundations that could scale

Cadence’s foundation layer unified tokens, accessibility standards, content guidance, and visual principles across products.

Teams could work from a common system and guidelines while preserving product-specific flexibility where business needs required it.

The components became a shared infrastructure

Cadence evolved into a reusable component ecosystem across multiple applications and business lines.

Components were designed, documented, versioned, and maintained as shared assets, reducing duplicate work and improving consistency across teams.

Documentation operationalized the system

Documentation became the primary interface between the system and the teams using it.

ZeroHeight and Storybook brought guidance, examples, release notes, onboarding, and support workflows into a shared source of truth.

As adoption grew, documentation became one of the system’s most valuable products.

Design-to-code pipeline

As the system matured, the focus shifted from creating assets to getting the pipeline and connections set up and working. Getting access to all the environments was also a task.

Design tokens, documentation, Storybook, a CI/CD workflow, and production code were linked into a shared system that reduced duplication and improved consistency across adopting teams.

The pipeline gave Cadence a path from design decisions to implementation, testing, release, and long-term maintenance.

Design-to-code architecture connecting design, documentation, Storybook, design tokens, and production implementation

Impact

Adoption grew across five business lines

Cadence expanded from a single-product initiative into a shared platform supporting multiple business lines, teams, and applications.

A common design language and implementation model allowed teams to build independently while maintaining consistency across the portfolio.

Bridged design and engineering at scale

Cadence and the UX Engineering function became foundational to frontend modernization.

The team supported Storybook adoption, token architecture, accessibility implementation, and React component development across multiple product teams.

Together, these investments turned Cadence into something engineering actually supported, not just a design-only resource.

Preparing For AI Workflows

The same foundations that supported design and engineering workflows also created opportunities for AI-assisted delivery.

Structured tokens, documentation, content guidance, and governance models became reusable operating context for tools like Figma Make, Claude, ChatGPT, and Azure-based workflows.

Components teach AI what to build, patterns teach it how experiences should work, and governance teaches it how decisions get made. Together, these files turn years of design, engineering, accessibility, and product knowledge into context that's just as useful to AI systems as it is to people.

Alignment was harder than implementation.
Michael Whitaker, on Cadence

Reflection

The work behind the work

Alignment was harder than implementation. The technical challenges were solvable; earning trust across design, engineering, and product is what actually determined whether Cadence succeeded.

The most important design decisions were not visual. Establishing governance, defining contribution models, building the business case for UX Engineering, and guiding the system through multiple reorganizations required as much design thinking as any component, token, or pattern.

Cadence taught me that design systems are organizational products as much as design ones. They succeed through shared ownership, education, trust, and a lot of ongoing stewardship.

Looking ahead, I believe the next evolution of design systems isn't additional components, it's infrastructure for AI.

The same foundations that made Cadence scalable (semantic tokens, governance, typed components, and design-to-code workflows) also prepare a system to support AI-assisted product development. The next step is making those systems something AI can actually consume, measure, and observe: expose them through machine-readable interfaces, evaluate generated UI against system standards automatically, and keep an eye on adoption, drift, and component health over time.