Deborah Foley Portfolio Deborah Foley Portfolio

Building design systems — and teaching AI to build them

Case Study 03 · Design Systems & AI

Building systems — and teaching AI to build them

For two decades, design systems have been my through-line. When the AI wave hit, everyone was chasing the same thing: fast, polished, interactive prototypes. They won early greenlights for my strategic proposals, too. But I looked under the hood: the plumbing was poor, and AI couldn’t map to systems. That’s what pushed me past the demo — to build and train the tooling that teaches AI to write quality, systematic code. This is both stories, in order: the systems I built at growing scale, and the machine I taught to build them.

Systems → AI

Two decades of building systems — and Alfred, the tool I built and trained to make AI build them right

Design systems at ecosystem scale, and the AI tooling that writes production-quality, systematic code — because good code isn’t subjective, and I know exactly how to specify it.

Outcome

The AI-assisted tooling and page-builders I built turn the design system’s structured docs and my own MD rule files into system-consistent generated code and faster prototypes — speeding delivery and lifting the team’s output on their protected growth time.

At a glance

Role

Strategic design leader and people manager — led Design Systems for Fire TV and Fire Tablet with design leads across the space, including 3P manufacturer HIG & customization support; and, with the AI tooling I built, building production code myself again

Scope

Design systems at ecosystem scale, plus the AI tooling that generates systematic, production-quality code

Also leading

Concurrently led the Amazon Appstore — the consumer storefront and the Amazon Developer platform — including its brand & marketing

Skills

Design-systems architecture · Governance · Front-end & AI tooling · Mentorship · Strategy · Stakeholder alignment · People management

On confidentiality. This work was done under NDA. The one shown artifact — the Alexa Human Interface Guidelines — is published publicly at developer.amazon.com/alexa. The rest is described without internal screens; full detail available on request.

A note on the through-line

From four Alexa device systems in 2022, to growing the next systems leaders, to enabling 130 third-party manufacturers in 2026 — and, when AI arrived, building the tool that finally makes it write production-quality code. One through-line, at ever-larger scale: systems. I’ve led them from both sides — as the developer who built them and the designer who governs them.

It started with an interactive-design degree, bridging design and code in Flash and ActionScript. When Flash was deprecated, that naturally moved me into front-end development — the era of responsive design and the first big frameworks, Bootstrap and jQuery. But I was already focused on lean, high-performance code and architecture, so I built my own component libraries and systems instead — before “design systems” was even a term. At Goldman Sachs I was hired to re-architect their wealth-management platform and build its component library — and it was my agile and scrum-master background that got both designers and developers to actually adopt it. By Chase, and then Amazon — one of the largest design orgs in the Amazon organization — I was strictly on the design side of governance, but always partnered closely with the systems-engineering teams. My role is strategy, architecture, and mentoring — setting the direction, governing the systems, and growing the teams who build them. Knowing both sides — the code and the design — is precisely what lets me teach AI to write good code and hold the bar.


2022 · Head of Design Systems, Alexa Devices

Four device systems, one atomic backbone

I led design systems for four device orgs — Fire TV, Echo, Fire Tablet, and Alexa Mobile — and turned a cleanup-and-spec team into system designers with real governance.

Alexa Human Interface Guidelines — published at developer.amazon.com/alexa
Alexa HIG · published

The problem I inherited

A cleanup team, not a systems team

The team was mostly doing production cleanup and writing dense spec docs — deeply technical, but not operating as system designers with governance. And the files fought them: 20–30 screen UX files with no grid and no auto-layout, so fixing one screen meant fixing all of them by hand.

The fix

An atomic architecture

I rebuilt the libraries on an atomic model — tokens → foundation → elements → patterns → templates — so a change propagates instead of being re-drawn everywhere, and so the team designs the system rather than cleaning up after it.

The library, and the 2022 arc

  1. Build

    Atomic libraries across four devices

    Foundation to templates, per device — Mobile, Tablet, Echo, and Fire TV each getting a real, governed system.

  2. Migrate

    Sketch → Figma, rebuilt atomically

    The org was on Sketch. To build a mature library I moved everyone to Figma, rebuilding every library atomically — each one by hand, from the ground up. Within a year we published a mature library.

  3. Partner

    An Echo-led unified strategy

    I partnered with the Echo-sponsored team writing a unified-system strategy across devices; a couple of my designers helped build their library too.

  4. Call it

    Fire TV builds its own

    I questioned the premise: Fire TV is a 10-foot, remote-driven screen — does it really unify with small touch devices? The team needed a system now, so we committed to Fire TV’s own.

  5. Deliver · Alexa HIG

    The published Alexa guidelines

    The unified thread led to the public Alexa Human Interface Guidelines at developer.amazon.com/alexa — which I helped consult, design, and deliver as part of the team. (Shown in the hero above.)

What the 2022 systems work delivered

Delivered

Component libraries · 4 devices

Built governed, atomic systems across the Alexa device family — Mobile, Tablet, Echo, and Fire TV.

Delivered

Sketch → Figma

Migrated the whole org, with the design-ops and training to make it stick.

Published

The first Alexa HIG

Internal and external guidelines — the one public artifact here.


The full journey

AI & Systems — the full journey

The full journey, shown with examples — available on request: Head of UX & mentorship at the Appstore; 1P/2P/3P enablement at ecosystem scale (130 manufacturers); and the AI journey — teaching AI to build on systems, the POCs, Alfred, the quality thesis, and the Fire TV refactor. Reach out for access. View more of my journey with examples →

Enablement & the AI journey

First-, second-, and third-party enablement at scale — and teaching AI to build on systems: the tooling, the POCs, and the quality thesis.

AI tooling — validating output against the rules

Now · Fire TV, refactored

Back in systems ownership, rebuilding Fire TV ground-up — built for humans and for AI.

Share Project :

  • Share Project :