Fragomen
Rethinking Case Work
Immigration Case Management Platform
Product Design Lead
I helped build Fragomen's first in-house UX team, hiring designers, establishing design standards, and leading research and product design for its immigration technology.
01
The Problem
Fragomen’s paralegal software was essentially a complex database administration tool. It was built around the structure of the underlying data rather than the natural workflow and mental models of the paralegals who used it every day.
A single paralegal could be responsible for more than 100 immigration cases at different stages, gathering information from applicants, corporate HR systems, legal documents, attorneys and other sources.
They manually entered and updated information piece by piece: what had been received, what was missing, who needed follow-up, where each case stood, and what needed to happen next.
Paralegals worked hard to meet the demands of the software. We wanted software that worked hard to meet the demands of paralegals.
02
Discovering the Human Glue
I shadowed and interviewed paralegals in Fragomen’s New York and Toronto offices to understand how immigration applications were actually getting done.
Much of their time was spent acting as human glue between disconnected systems and processes — requesting information and documents through email, tracking responses, following up with applicants, reconciling information and manually updating case status.
I also noticed spreadsheets everywhere. Paralegals were building makeshift dashboards outside the product to understand their own workload.
The spreadsheets were doing what the software should have done. That was an important signal.
03
Building a Case Work Engine
Once I understood how paralegals actually managed their workload, I designed a new experience organized around cases, status and required actions rather than database fields.
The dashboard reflected the questions paralegals were already asking themselves: what is complete, what is missing, which cases are blocked, what requires my attention next.
But reorganizing the interface was only part of the opportunity. If the system knew an applicant needed to provide information, why should a paralegal have to manually send the request? If the applicant didn’t respond, why should the paralegal have to remember to follow up?
The system could take greater responsibility for routine coordination by automatically requesting information, tracking responses, sending reminders, and communicating status.
04
Freeing Humans to Do What They Do Best
This shifted the product from helping paralegals track administrative work toward actually doing some of that work for them.
The goal wasn’t to automate the judgment that made paralegals valuable. It was to automate the predictable tasks that consumed their time — freeing them to focus on client relationships, legal judgments, and complex exceptions.
Working closely with Product and Engineering, I defined the workflows, information requirements and automation needed to support immigration cases. The resulting initiative became known internally as “Big Bang.”
It represented a fundamental shift in the product: from software organized around immigration data to software organized around getting immigration work done.
Outcome
Projected 70-80% reduction
In manual processing
Our analysis of the redesigned workflow projected a 70 to 80 percent reduction in manual processing.
The largest gains didn’t come from making data entry faster. They came from removing predictable administrative work that didn’t require human judgment.
The software no longer just stored the work. It could begin helping get the work done.
