01, Project Overview
EY Data Fusion
Enterprise Data Platform
Data Fusion is EY's cloud-native ETL platform, built on Databricks, designed to handle data preparation, governance, quality and AI-powered data exploration for financial services and enterprise clients. The platform combines structured data pipelines with GenAI features to make data work faster and more reliable for large organizations.
EY Data Fusion Platform Overview, click to watch on YouTube
My role: Lead UX/UI Designer at EY, responsible for the end to end design of the Data Fusion platform. I own design across all modules, work directly with engineering on implementation, and run UX reviews to keep the product consistent with what was designed.
Enterprise UXUI DesignDesign SystemsETL PlatformGenAI FeaturesFigmaMotifScrum
Managing Director
Product ownership
3 Tech Leads
1 AI focused, 2 product focused
Product Manager
Roadmap and priorities
Product Owner
Requirements and backlog
Scrum Master
Sprint and ceremony management
Engineering Team
Frontend, backend, AI
02, Process
How decisions get made
without direct user access
Data Fusion has no direct line to end users for usability testing. Instead, the platform relies on a structured internal process combining QA findings, customer support feedback and stakeholder alignment to decide what gets built each quarter.
1
QA finds issues and gaps. A dedicated QA team continuously tests the platform and surfaces bugs, inconsistencies and improvement opportunities.
2
Customer support tickets get reviewed. Feedback collected through support tickets is analyzed alongside QA findings to spot patterns in real user frustration.
3
Leadership aligns on priorities. The Managing Director works with the 3 Tech Leads, the PM and the PO to review all findings and decide what matters most.
4
PI Planning sets the quarter. Priorities turn into a quarterly plan covering both new features and improvements to existing ones.
5
User stories and sprints. Work is broken into user stories and run through Scrum sprints, with daily standups involving stakeholders, PM, devs and tech leads.
6
UX review closes the loop. I check every shipped feature against the original design and flag inconsistencies before they reach production.
03, What I Designed
Features shipped
over the past year
Below is a direct breakdown of the features and improvements I designed across the platform, organized by module.
AI and Data Explorer
SQL Assistant for Data Explorer, AI-powered query support
Data Explorer AI Chat experience
GenAI chart and table generation concepts
Dark Mode for the SQL Assistant
Databricks-style querying improvements
Chatbot feedback capture flows
Mapping and Data Transformation
Source Node Mapping name selection
Raw to CDM Mapping UI across 5 sprints
JSON support in mapping UI
Mandatory column validation for cleanse flows
Dynamic mapping tabs for readability
CDM Data Dictionary integration
PII, PCI and Governance
PII/PCI onboarding module redesign
Profiling module for sensitive data
PII tag propagation across CDM layers
Element tag creation for tenant rules
LLM-based anomaly detection design
Data Quality and Monitoring
DQ Insights with interactive dial-to-graph behavior
Anomaly detection in Outliers tab
Batch job notifications system
Read-only access restrictions across modules
Platform Consistency and UX
Localized date/time display across all modules
UI scaling for laptop users at 150%
Pagination and scrolling improvements in Data Profile
Data Fusion sitemap for navigation
Global navigation redesign
Micro frontend integration design
Access and Onboarding
Access Request modal and email flow redesign
Onboarding and mapping user journeys
Project admin creation UI for WAM apps
Design System
Migrated COA module to the updated Motif Design System
High-fidelity Figma designs delivered across all modules
Consistency maintained across Mapping, Data Explorer, DQ and Onboarding
04, Working Method
Design system
and daily collaboration
Every screen is built using Motif, EY's internal design system, which keeps the platform consistent as new features ship every sprint. Daily meetings with stakeholders, PM, developers and tech leads keep design decisions grounded in real technical and business constraints.
8+
Platform modules covered
Daily
Stakeholder alignment meetings
Motif
Internal design system used
Quarterly
PI Planning cycles
05, Conclusion
Designing without direct
user access
Data Fusion is an example of designing at scale inside a large enterprise platform, where direct user research is not always possible. The work depends on triangulating QA findings, support tickets and stakeholder input to make informed design decisions, then validating through UX review once features ship.
Working closely with 3 Tech Leads and engineering kept every design technically grounded before it reached production
Using Motif consistently across 8+ modules kept the platform coherent despite constant feature additions
Running UX review after every release caught inconsistencies before they reached users
"Designing Data Fusion means working without direct access to the people who use it every day. The discipline comes from listening carefully to QA, support tickets and stakeholders, and staying consistent through the design system no matter how fast the platform grows."
Otávio Bonato, Lead UX/UI Designer, EY