Overview
The dashboard every immigration officer
dreaded opening.
ELIS, the Electronic Immigration System, is used by USCIS adjudicators, supervisors, and clerks to process hundreds of thousands of immigration applications annually. My Cases was the first screen officers saw every morning. It was broken in almost every functional way.
I came in as the sole production designer. Over 7 months I led everything: stakeholder alignment, generative research, 7 sprint facilitation, in-person usability testing at Vermont Service Center, and mentoring junior PMs who had never run a structured design process.
The Problem
Five systemic failures
in one dashboard.
Before touching a single wireframe, I needed to understand why the existing system was failing, not just what it looked like. Through heuristic evaluation and initial stakeholder interviews, five critical failure modes emerged:
No prioritization signals
Cases appeared in an undifferentiated list with no urgency indicators, deadline tracking, or visual hierarchy. Officers had to manually sort through hundreds of cases to identify what was urgent.
Zero customization
A fixed set of data columns, including many irrelevant to a given officer's role, that couldn't be hidden, reordered, or saved. Adjudicators and clerks saw identical interfaces despite very different workflows.
Single-case work queue
The legacy "Get Work" flow only allowed pulling one case at a time from a single queue, a catastrophic bottleneck in a high-volume processing environment.
No organization system
With hundreds of open cases, officers had no way to group, bookmark, or follow-up tag cases. Everything lived in one flat, unsortable list.
Broken global search
Search returned inconsistent results that made it harder, not easier, to locate specific cases. Officers had resorted to browser-native Ctrl+F as a workaround.
The real cost
Slow adjudication times aren't just a UX problem, they directly impact the backlog that determines how long real people wait for immigration decisions. The stakes were unusually high.
Before state, legacy My Cases workflow and interface
My Role
Sole production designer.
Every layer.
Being the only production-level designer on a government project meant operating at every altitude simultaneously. I wasn't handed a backlog and told to design, I built the process from the ground up.
What I owned
4-month design plan · 3-month launch/adoption roadmap · 7 design sprints · 15 user interviews · 12 in-person usability sessions at Vermont Service Center · 72-participant testing program · Mentoring junior PMs · Stakeholder alignment · 70 approved user stories · 6 shipped features
One thing I'm particularly proud of: the mentoring dimension. Several of the junior PMs on the project had never worked with a designer who ran structured sprints and involved users at every stage. Part of my job became teaching them what a design process could look like, and getting them invested in the outcome as a result.
Research
You can't design for people
you haven't met.
I started with 4 SME interviews to understand technical constraints, then a heuristic evaluation of the existing interface. From there: 60+ user stories and a recruiting screener targeting three distinct personas.
⚖️
Adjudicator
Decides on immigration benefits. Manages complex, competing case details. Needs to assess case urgency quickly and make high-stakes decisions daily.
📋
Supervisor
Manages field office task queues. Overwhelmed by shifting priorities and unpredictable backlogs. Needs visibility into team workload at a glance.
📁
Clerk
Handles document preparation and intake. Challenged by locating and organizing case materials across an unstructured system.
One quote from an adjudicator that became the design north star: "I know it's in there somewhere, I just don't know where."
Design Process
Seven sprints.
Every feature earned.
I organized 7 design sprints over 6 weeks, each targeting a specific feature area of the My Cases experience. Every sprint followed the same cycle: internal discovery → design → test → synthesize → iterate. No feature moved forward without user validation.
| Sprint |
Focus Area |
Key Output |
| 01 | Simple Tables & Basic Structure | Base table architecture, sortable columns |
| 02 | Editing & Alerts | Inline editing states, status alert system |
| 03 | Filtering & Bulk Actions | Multi-criteria filter panel, bulk case selection |
| 04 | Customizable Columns & Quick View | Column config modal, side-panel Quick View |
| 05 | Expandable Tables | In-row case expansion, nested detail views |
| 06 | Global Search | Cross-entity search with consistent result types |
| 07 | My Folders | Custom folder system, drag-and-drop case organization |
After the sprint series I conducted 2 rounds of in-person usability testing at Vermont Service Center with 12 participants. Being physically on-site was a deliberate choice, watching officers interact with the prototypes in their actual work environment revealed friction points that remote testing would have missed entirely. I made same-day iterations between sessions.
Key Design Decisions
The decisions that
moved the numbers.
Decision 01, Customizable Columns
Officers were drowning in irrelevant data. I designed a column customization system letting users show/hide and reorder up to 8 columns with saved preferences per user. The key insight: different roles needed radically different data views. One interface was never going to serve all three personas.
Decision 02, Quick View Side Panel
Officers needed to assess a case without losing their queue position. The side-panel Quick View surfaces the 6 most critical data points inline, eliminating the open/close loop that was adding seconds to every case review and multiplying across hundreds of daily interactions.
Decision 03, Get Work Redesign
This single change had the most measurable throughput impact. The old system: one case at a time, one queue. The new system: bulk selection from multiple queues simultaneously, plus a "Search for Available Work" tab for manual assignment. The 35% adjudication time reduction traces directly here.
Decision 04, My Folders
A bookmarking system allowing officers to create up to 10 custom folders and drag-and-drop cases into them. Simple to describe, significant in impact, for the first time, officers had personal organization inside a system that had always treated their work as a shared, undifferentiated pile.
Final Designs
Five features.
One cohesive system.
Each feature below was independently validated in usability testing before being assembled into the final shipped interface.
Customize Columns modal, users can show/hide and reorder up to 8 columns, saved as default per user
Quick View panel open, 6 critical case data points surfaced inline without navigating away from the queue
Get Work modal, multi-queue selection with "# Cases to Pull" per queue. The single feature most directly responsible for the 35% adjudication time reduction.
My Folders modal, up to 10 named folders, color-coded and reorderable. The first personal organization system in ELIS history.
Global Search, Alien Number match surfaced inline with case summary. No page navigation required.
Accessibility
Designed for
every officer.
USCIS platforms are used by a diverse workforce with varying visual abilities. I incorporated a 12-color accessible palette adapted for color blindness, ensuring status indicators, folder colors, and priority signals were distinguishable across all four major vision deficiency types.
12-Color Accessible Palette, tested for Deuteranopia, Protanopia, Tritanopia, and Normal vision
Normal
Deuteranopia
Protanopia
Tritanopia
Each color remains distinguishable across all four vision types, ensuring folder indicators, status badges, and priority signals are universally readable.
Outcomes
Numbers that
surprised everyone.
The adoption target was 50% within 4 months. The adjudication target was a 25% reduction. We didn't just hit them, we significantly exceeded both.
| Metric | Target | Result |
|---|
| User adoption rate (4 months post-launch) | 50% | 98% |
| Adjudication time reduction | 25% | 35% |
| Total research participants | — | 72 |
| Task completion rate (usability testing) | — | 90% |
| User stories approved | — | 70 |
| Features shipped to production | — | 6 |
The 98% adoption rate came from bringing users into every sprint and making them feel ownership over the outcome. The project also won a collaboration award, specifically called out for an inclusive process that made users feel heard.
Reflection
What this project
taught me.
Takeaway
"Working as the sole designer on a lean government team taught me how to operate at every level simultaneously. The constraint forced me to build process, not just deliverables. The 98% adoption rate came from bringing users into every sprint..not just from the quality of the designs."
