Downloading 50 pay stubs shouldn't take 12 minutes, but in a system where data is siloed and APIs aren't built for bulk, that was the reality. My task was to move the needle from weeks of manual labor to minutes, without compromising the strict compliance standards required by law.
01 - Context & Background
Two types of practitioners rely on it daily : COS (customer service) teams managing payroll on behalf of clients, and client HR administrators running their own. When either group needs pay stubs for legal, audit, or offboarding purposes, the stakes are immediate and the margin for error is zero.
The feature needed to achieve parity with legacy platform, which already supported bulk download. The gap wasn't just a UX inconvenience, it was leaving COS teams exposed to weeks of manual labor for every legal or DOL audit request.
02 - The Problem to Solve
When a client received a DOL audit notice or an employee filed a lawsuit, a COS practitioner would open iReports, search through a PDF that could contain 25,000 employees, extract individual pages, and repeat for every pay period across years of history. Modern application improved the interface but not the loop.
And the compliance pressure was real. Incorrect pay statements don't just create operational delays in states like New York, they create legal exposure. A wrong statement can end in a lawsuit. That's the environment in which this feature needed to work flawlessly.
"It used to take a week at least when data was older. Imagine if they asked for a longer period, then data's not in iReports, you have to extract it from archives, and then again you go through one file at a time, look into all the pages."
03 - My Role & Collaboration
I led end-to-end UX design for this feature, from facilitation of the Design Thinking workshop through research synthesis, concept exploration, and final UI. The hardest design work wasn't the screens. It was navigating the tension between what practitioners needed, what compliance required, and what the iPay API could actually do.
My responsibilities spanned: mapping and critiquing the current-state workflow, leading empathy exercises with COS practitioners, creating three design concepts for critique, user testing, and synthesising findings into a simplified final direction that diverged from my initial hypothesis.
04 - Discovery & Research
The Design Thinking workshop ran for 90 minutes with five participants across compliance, COS operations, and product. The most important insights weren't in the initial problem statement - they emerged from the debate.
05 - Three Concepts Explored
The workshop generated three distinct design directions. Each was sketched, critiqued in group session, and ultimately tested against real users. The process of elimination was as important as the final direction.
My initial hypothesis : that a unified modal combining bulk pay statement downloads with pay summaries would be a 'one-stop shop', failed during usability testing. Users were overwhelmed by the mixing of download actions and data viewing. That 'failure' was essential: it proved that the internal (payroll) and external (pay statement) workflows were fundamentally different, confirming that our design needed to be de-coupled to meet both user needs and compliance standards.
06 - The Final Solution
Two separate tabs, focused experiences, each optimised for its distinct use case, turned out to work better than any unified approach.
The key design principle wasn't "put everything in one place." It was "make the 80% use case invisible." For Pay statement bulk downloads: filter, select all, download. For Pay summary statement : filter, select all, view Quick Totals. A focused modal showing aggregate figures from selected periods, with options to export.
07 - Challenges & Concepts Not Pursued
Three of the most important design decisions in this project were decisions not to build something. Each discarded direction taught something that improved the final solution.
The initial concept bundled bulk downloads, Quick Totals, and detailed breakdown into a single modal with tabs, plus configuration options for SSN masking and YTD toggles. Round one user testing was unambiguous: "Is this for downloading or just viewing data?" Configuration options felt overwhelming for what users described as a quick task. We separated the use cases entirely and removed redundant controls.
The majority of workshop participants wanted the two tabs merged into one. The technical reality : different data sources, different legal treatment, and upcoming tax statements requirements, made this premature for MVP. The two-tab approach was kept with a documented roadmap for consolidation pending engineering assessment.
Dynamic Right Panel concept failed to survive the realities of our design system’s governance and the high engineering cost of building unsupported, non-standard patterns. Instead of forcing a novel UI, we pivoted to scalable, platform-consistent solutions that prioritized maintainability and a shippable experience.
08 - Impact & Outcomes
The metrics were not aspirational targets set at the beginning of the project. They were post-launch measurements taken after two weeks of live use. The numbers validated that separating use cases had been the right call.
09 - Reflections & Learnings
Our initial design was cluttered with redundant toggles for SSN masking, YTD options, and formatting controls that simply replicated rules already enforced by policy or law. This taught me that when you find yourself piling on settings, you’re failing to leverage smart defaults; design shouldn't ask users to configure what the system should handle automatically.
I tested the complex unified modal in round one because I assumed simpler flows needed less validation. They don't. They need it more, because the design has fewer places to hide when assumptions fail. Testing a simple multi-select → download flow first would have established a baseline faster and led to the pivot earlier.
The final design looks obvious in hindsight, but it was forged through rigorous testing, discarded concepts, and the deliberate decoupling of complex workflows. Simple interfaces aren't the absence of complexity; they are the byproduct of difficult decisions.