Employee Mobility Platform
Uber for Business — 2024
Overview
The Employee Mobility platform enables organizations to manage transportation benefits for their workforce. This project focused on simplifying complex admin workflows and improving the configuration experience for enterprise customers.
What was the core design challenge?
Enterprise admins struggled with a fragmented interface that required too many steps to configure basic mobility programs.
- Low task completion rates on core admin workflows
- High support ticket volume from configuration confusion
- Multiple disjointed screens for a single setup flow
How did user research shape the solution?
Interviews with 12 enterprise admins revealed that the primary pain point was context-switching between fragmented configuration screens.
- Admins needed a single-flow experience for program setup
- Preview capabilities were essential for confidence before publishing
- Bulk operations were underserved but highly demanded
Some details have been simplified for public presentation. Internal metrics and exact feature names have been generalized.
Problem
Enterprise admins at Uber for Business struggled with a fragmented, multi-step configuration interface for employee mobility programs. Task completion rates were low, support tickets were high, and new admins required extensive onboarding.
Solution
A unified configuration flow with real-time preview, intelligent defaults, and progressive disclosure. Replaced 6 separate screens with a single guided setup experience.
Impact
40% reduction in admin setup time, 60% fewer support tickets, and 3 design system components adopted across teams.
Context & Constraints
Uber for Business serves thousands of enterprise customers managing employee transportation programs. The existing admin platform had been built incrementally over 3 years, resulting in fragmented workflows across multiple disconnected screens.
The platform needed to support organizations ranging from 50-employee startups to 100,000-employee enterprises, each with unique configuration needs including geographic policies, approval workflows, and budget controls.
Constraints
- TechnicalMust integrate with existing Uber internal APIs without breaking backward compatibility for current enterprise clients.
- BusinessCannot disrupt active customer accounts. Migration from old UI to new UI must be seamless with zero downtime.
- Timeline12-month delivery window with phased rollout. MVP needed within first 4 months for internal testing.
Assumptions
- Enterprise admins are semi-technical and familiar with SaaS admin panels
- Most admins configure programs during initial onboarding, then make infrequent edits
- Mobile admin access is not a priority — 95% of admin tasks happen on desktop
Goals & Success Criteria
Reduce program setup time
Lower support burden
Improve admin confidence
Hypotheses
- A unified configuration flow will reduce context-switching and improve completion rates
- Real-time preview will increase admin confidence and reduce "publish anxiety"
- Intelligent defaults will eliminate 60% of required configuration decisions for standard programs
My Role
As the sole designer on this project, I owned the complete design lifecycle. I collaborated closely with the PM to define scope and priorities, worked with the UX researcher on study design and synthesis, and partnered with engineering throughout implementation.
Responsibilities
- Led end-to-end design from discovery through launch
- Conducted user research including 12 admin interviews and 3 contextual inquiries
- Created and maintained the design system components for the admin platform
- Facilitated design reviews and cross-functional alignment sessions
- Partnered with engineering on implementation details and design QA
Collaborators
- Aditya Raman—Product Manager
- Sarah Chen—Engineering Lead
- Priya Nair—UX Researcher
Project Timeline
Discovery & Research
Conducted stakeholder interviews, admin shadowing sessions, and competitive analysis.
Problem Framing
Synthesized research into opportunity areas. Defined design principles and success metrics.
Concept Exploration
Explored 3 design directions. Validated concepts through rapid prototyping and stakeholder feedback.
Detailed Design & Testing
Built high-fidelity prototypes. Conducted 2 rounds of usability testing with 8 enterprise admins.
Engineering Handoff & Build
Detailed specs, component documentation, and weekly design-engineering syncs during build phase.
Launch & Measurement
Phased rollout to 20%, then 50%, then 100% of enterprise accounts. Tracked key metrics post-launch.
Discovery & Stakeholder Mapping
I started by mapping the stakeholder landscape and understanding who touches the admin platform. This revealed a more complex ecosystem than initially assumed — not just "admins" but travel managers, finance controllers, HR coordinators, and IT administrators each had distinct needs.
Key Insights
- Six distinct admin persona types use the platform with different task frequencies
- Travel managers were the most frequent users but had the least configuration power
- Finance controllers needed reporting integration that the current UI completely lacked
Decision
Focus on the "Program Admin" persona (70% of users) while ensuring extensibility for other roles.
Rationale
Program admins perform the most complex and frequent configuration tasks. Optimizing for them maximizes impact while their workflows encompass most of what other roles need.
Information Architecture Redesign
The existing IA spread program configuration across 6 separate screens with no clear hierarchy. Through card sorting with 8 admins and tree testing with 15 participants, I developed a new IA that grouped related configurations into logical stages.
Key Insights
- Admins expected a linear setup flow similar to onboarding wizards in other SaaS tools
- The concept of "program" was overloaded — admins conflated policies, budgets, and riders
- Card sorting revealed 4 natural groupings: People, Policies, Budget, and Settings
Decision
Adopt a stepped wizard pattern for initial setup with a dashboard for ongoing management.
Rationale
The wizard pattern matches admin mental models from other enterprise tools and naturally guides them through the required configuration sequence.
Visual Design & Component System
I designed the visual language to feel professional yet approachable. Enterprise admin tools often feel sterile — I wanted to inject warmth through typography, spacing, and micro-interactions without sacrificing information density.
Key Insights
- Admins spend 4-6 hours daily in the platform — visual fatigue is a real concern
- Dense information displays were preferred over simplified views for experienced users
- Color coding for program status reduced scanning time by ~30% in testing
Decision
Build 12 new design system components that became shared across 3 product teams.
Rationale
Investing in shared components paid for itself within 2 sprints. Other teams could adopt the same patterns, ensuring consistency across the entire admin experience.
Research
Methods
Participants
12 enterprise admins, 8 internal stakeholdersFindings
Context-switching was the #1 pain point
10 of 12 admins mentioned losing their place when moving between configuration screens.
Preview anxiety was universal
All admins expressed fear of publishing incorrect configurations. 7 of 12 used a staging environment as a workaround.
Bulk operations were a top unmet need
8 of 12 admins manually repeated the same configuration across multiple programs, sometimes spending hours on repetitive tasks.
Synthesis
The platform has a trust problem, not a feature problem
Admins had the tools they needed but lacked confidence in using them. Preview and undo capabilities were more important than new features.
Configuration complexity scales with organization size
Small orgs need simplicity, large orgs need power. One-size-fits-all UI fails both groups.
Admin expertise varies wildly within the same account
A single enterprise account might have a power-user travel manager and a brand-new HR coordinator using the same interface.
Frameworks
The synthesis phase revealed that the core problem was not about features but about confidence and flow. Admins had the capabilities they needed but struggled with the fragmented interface and lack of feedback.
Ideation
Wizard Pattern
A guided step-by-step flow that walks admins through program configuration in a linear sequence. Familiar pattern from other SaaS tools.
Hub-and-Spoke Dashboard
A central dashboard with deep-linked configuration panels. Supports non-linear exploration but requires more cognitive load.
Hybrid Approach
Wizard for initial setup + dashboard for ongoing management. Best of both worlds but more complex to build.
Selection Criteria
We evaluated each approach against 4 criteria: first-time setup speed, ongoing management efficiency, development effort, and scalability for future features. The hybrid approach won on all dimensions except development effort, but the long-term benefits justified the investment.
Prototyping
Hi-fiI built a high-fidelity prototype in Figma that covered the complete setup wizard flow (4 steps, 12 screens) and the management dashboard. The prototype included realistic data, edge cases, and error states to ensure comprehensive testing.
Interactive prototypes were shared with the engineering team early to identify technical constraints before finalizing the design.
Usability Testing
Tasks & Results
Configure a new mobility program for 500 employees
7 of 8 completed successfully. One participant struggled with the geographic policy step due to unfamiliar terminology.
Edit an existing program budget allocation
All participants completed within 2 minutes. Dashboard layout made this task significantly faster than the old interface.
Preview and publish program changes
The preview feature was universally praised. Participants reported feeling much more confident about publishing changes.
Recommendations
- Simplify geographic policy terminology (replace "geofence" with "area")
- Add inline help tooltips for less common configuration options
- Provide a "duplicate program" feature for admins managing similar programs across regions
Outcome & Impact
Average time to configure a new mobility program dropped from 25 to 15 minutes.
Configuration-related support tickets decreased within 3 months of launch.
Up from 62% on the previous interface.
New components adopted by 3 other product teams.
Qualitative Outcomes
- Enterprise customer satisfaction scores improved by 12 points post-launch
- Internal support team reported significantly reduced escalation volume
- Two competitor analyses cited the new admin experience as an industry benchmark
The redesigned platform exceeded all three success criteria established during the goals phase. The combination of a guided setup flow, real-time preview, and intelligent defaults created a significant measurable improvement in admin efficiency and confidence.
Reflection & Learnings
What Worked
- Early and continuous research involvement kept the design grounded in real admin needs
- The hybrid wizard + dashboard approach scaled well from MVP to full release
- Building shared design system components created value beyond the immediate project
What Didn't
- Initial stakeholder alignment took longer than expected — should have started with a formal design brief
- The first prototype was too polished, making stakeholders resistant to fundamental changes
- We underestimated the migration complexity for existing customers
Key Learnings
- Start with rough wireframes in stakeholder reviews — polish signals finality
- Enterprise admin tools need both a "guided" and "expert" mode from day one
- Measuring time-to-task is more meaningful than satisfaction scores for admin tools
Next Steps
Bulk operations for multi-program management
highEnable admins to apply configuration changes across multiple programs simultaneously. Highest-requested feature from enterprise accounts.
Role-based dashboard views
mediumCustomize the dashboard layout based on admin role (travel manager vs. finance controller vs. IT admin).
AI-powered program recommendations
lowSuggest optimal program configurations based on organization size, industry, and usage patterns.
The next phase focuses on scaling the platform for larger enterprise accounts and introducing intelligent automation to reduce manual configuration effort.