Design Case Study: Platform Unification

Unifying an enterprise platform under one design system

Role-based information architecture and design-engineering pairing to bring an Enterprise Platform (Self-Hosted) onto a shared design system.

My Role: Lead UX Designer (Staff UX Designer)

Timeline: May to November 2025

Team: Product Manager, Engineering Manager, Frontend and Backend Engineers

The platform before and after the redesign, side by side

At a Glance

15 6 • Left Navigation Items

Consolidated and now grouped by category

13 6 • Subnavigation Tabs

Reduced with no functionality lost

10+ 1 • Disparate Patterns

Consolidated patterns into a single design system

2 • Role-Based IAs

Created specifically for Admins and Practitioners

The Problem

Built without UX, the product had outgrown its interface:

15 navigation items with no hierarchy
13 tabs on a single details page
Red "manage" labels on nearly every row (though red signals danger)

It also looked and behaved differently from the company’s other products, including their desktop app.

The Opportunity

In 2025, a new Chief Product and Technology Officer made cross-product UX alignment a priority for the first time.

This product went first, making it the proving ground for the design system inside a complex enterprise product.

My Role

Lead UX Designer, embedded daily with product and engineering.

I owned the IA and design direction, facilitated key decisions, and paired with the team's lead senior front-end engineer.

The Users

Admins: IT and security staff who govern the platform.

Practitioners: data scientists and developers who use it every day.

One crowded navigation served both. Separating their experiences was an important principle of the redesign.

The original interface with one long left navigation — Before: the original interface, with admin, account, and daily-work items in one long left navigation.

Constraints

No new features: align and restructure without changing what the product did.
A fixed deadline: about six months to a planned November 2025 soft launch.
Legacy code and tech debt, with limited engineering capacity.
A design system still being built, so some components had to be created along the way.

Step 1

I split one crowded navigation into two role-based structures

I designed one information architecture for Admins and one for Practitioners, mapping the pages, tabs, and actions each role needed.

Navigation went from 15 items to 6 grouped by category, and admin and account tools moved to a new user menu in the upper right. The IA also planned ahead for the product's new AI/ML areas.

Admin information architecture diagram — Admin IA, including the new upper-right menu for admin and user settings

Practitioner information architecture diagram — Practitioner (IC) IA, with only the pages and actions daily work requires.

Before

After

Step 2

In one live workshop, the team cut 13 tabs to 6

I facilitated the Engineering Manager, two Product Managers, and several engineers through every tab on the channel detail page, capturing decisions in Miro.

The PMs' customer and usage data grounded each decision.

Any removed tabs moved to better homes and a few were merged, so nothing was lost.

Before · 13 tabs

After · 6 tabs

I chose a workshop over providing a design because the engineers knew the codebase and the PMs knew the customers. Deciding together meant the team owned the result.

Step 3

Twice-weekly pairing let us fix inconsistencies at the pattern level

I proposed a pairing practice with the team's lead senior front-end engineer.

Twice a week, she brought screenshots of patterns that didn't match the design system, with her proposed fixes, and we decided together directly in staging.

Before

After

10+ 1 • Disparate Modals

In one session, we turned at least 10 modal patterns into one consistent pattern.

Added to the design system

That modal joined the shared design system, so every product team could use it.

Color by meaning

Red was reserved for destructive actions, and the rest of the interface moved to calm, neutral tones.

Step 4

A mid-project walkthrough built support at the top

In August 2025, I demoed the redesign live for the CEO and the VPs of Engineering and Product, who responded positively to the progress.

Outcomes

Finished on time and ready to ship. The full redesign was complete for the planned November 2025 soft launch, within scope and despite legacy code and limited capacity. Its release was tied to a new AI/ML section that didn't yet have customers, and in spring 2026, acquisitions shifted company priorities and halted the project.

What I Learned

Pairing is a real way to get design work done. Deciding together in the live product meant I didn't need full Figma specs for every screen. We moved faster, avoided handoff gaps, and the engineer became a partner in holding the quality bar.

Next time, I'd push for direct access to customer data, so design can interpret it firsthand.

Testimonial

"She asked engineers real questions about the system, then used those answers to design experiences that were both delightful and buildable. She knew how to scope, how to ship something valuable now and make it better next iteration, rather than trying to boil the ocean."

— Engineering Manager, Web Experience team