Mathews MM

Global Automotive Client · UST · 2021–2025

Enterprise Data Catalogue UX

A four-year engagement. Started with a heuristic audit of a broken system, ended with a live design system used across multiple modules and release cycles.

Enterprise UXDesign SystemsData GovernanceIARole-based Access

Overview

At UST, I was the primary UX designer on a data catalogue platform built for a global automotive manufacturer's internal teams. The client needed a single, trusted place where analysts, engineers, and data stewards could find, understand, and request access to datasets. What existed instead was a fragmented mix of Excel files, SharePoint folders, and legacy tooling that nobody fully trusted.

Over four years I worked across the full design cycle: auditing the existing system, shaping the information architecture, designing core workflows, building a design system, and supporting engineering through multiple release cycles. The platform grew from a basic repository into a governed catalogue used across business units and user roles.

Engagement

4 years · multi-release

My role

Lead UX · sole designer

Platform

Internal enterprise web app

Interface patterns

Anonymised UI representations, actual screens are under NDA. These illustrate the interaction patterns and component decisions.

Dataset Search, anonymised representation

Domain
Owner
Status
Classification
FinanceActiveRestricted
Supply ChainActiveOpen
HRReviewRestricted

Dataset Detail View, anonymised representation

Owner

Domain

Classification

Last updated

Source system

Status

Description

Access Request & Approval Flow, anonymised representation

Request submitted
Steward review
3
Manager approval
4
Access granted
!

Design System, component patterns

Status badges

ActivePendingRestrictedDeprecatedIn Review

Empty state

Filter bar pattern

All
Finance
Supply Chain
HR
IT

The problem

01

Scattered data landscape

Dataset definitions, ownership records, and access policies lived across Excel, SharePoint, and internal wikis. Finding a dataset, let alone understanding it, could take days.

02

No governance UX

There was no structured way to request access to sensitive data or track approvals. Processes happened over email and chat, creating audit gaps and inconsistent access control.

03

A tool built by engineers, for engineers

The existing repository was technically functional but assumed deep system knowledge. Business users, including analysts, product teams, and domain leads, found it unusable.

Process

01

Audit & Research

Heuristic review of the developer-built tool and existing Excel-based workflows. Workshops with data stewards, engineers, and managers to surface pain points and turn them into user stories.

02

IA & Workflows

Defined a layered IA around datasets, domains, lifecycle states, and roles. Established core patterns for search, review, request, and approval before touching the UI.

03

Design System

Built a component library covering tables, filters, forms, status indicators, and empty states. The goal was consistency across modules without requiring design to be involved in every build decision.

04

High-Fidelity UI

Designed search, dataset detail, access request, and approval flows for three distinct user types: data consumers, stewards, and managers.

05

Implementation Support

Partnered with dev teams through sprints, refining edge cases, reviewing builds, and keeping the system coherent as the product evolved across multiple releases.

Key design decisions

Three decisions that shaped the platform's direction more than anything else.

IA around user goals, not system structure

The existing tool was organised around how the database was structured internally. We reorganised around what users actually needed to do: find a dataset, understand it, request access, or manage governance. That meant domain-based navigation, lifecycle states, and role-filtered views instead of a flat list.

Single approval workflow for all sensitive data

Every team had their own informal process for access requests. We designed a unified request and approval flow with clear states, notifications, and audit trails. The hard part was making governance feel like a natural step in the workflow rather than a bureaucratic gate bolted on afterwards.

Design system before screens

With multiple modules in scope and dev teams working in parallel, I prioritised establishing core patterns first: tables, filters, status indicators, empty states, forms. This prevented fragmentation and made every new module faster to design and build.

Impact & reflection

The platform evolved from a developer-built repository into a governed, multi-module catalogue used across business units. The design system reduced fragmentation as new modules were added, and the approval workflow replaced ad-hoc email processes with a structured, auditable flow.

The biggest lesson from four years on this: enterprise UX is as much about systems thinking and stakeholder alignment as it is about screen design. The best decisions I made were the ones that reflected how the organisation actually worked, not how I assumed it should.

More case studies