Accessibility Remediation Across Six Enterprise Healthcare Platforms

Context & Goals

The client is a major enterprise healthcare software company serving hospital networks, workforce management teams, and clinical operations. For enterprise healthcare, accessibility is not optional: WCAG 2.2 and Section 508 are tied to federal procurement obligations and enterprise contracts, and falling short is a legal and commercial risk. They engaged Level Access, an independent auditing firm combining automated scanning with manual expert testing, to audit six platforms: Workforce Management Android mobile, Workforce Management iOS mobile, Workforce Management desktop, Provider Management Platform, Provider Data Management Hub, and Clinical Evidence AI for Payers. The goal was to resolve all confirmed findings and leave each team with a process they could sustain independently.

Constraints

Six product teams operating independently, each with its own sprint cadence, familiarity with accessibility, and tolerance for design changes touching shared components. No shared process, no unified stakeholder, no single way of communicating findings that worked across all six teams. Every platform required its own working relationship and its own negotiation of scope.

Process

Audit review and validation
Before remediation began, I reviewed the full audit output directly in Level Access, resolving duplicates, unclear findings, and errors. This reduced the raw audit to 1,274 confirmed, actionable findings before any design work.

Organizing the work
Full engagement managed in Asana with finding category summaries and weekly status updates for managers. Per-platform spreadsheets for developers with responsible discipline, WCAG criterion, and Figma links. Figma files organized by platform screen rather than finding type.

Finding prioritization
Systemic findings addressed at the design system level. One-off issues handled platform by platform in order of severity and retesting schedule.

Remediation and retesting cycle
Remediate, validate against Level Access audit, implement, formal Level Access retest. Repeated across all 1,274 findings. I ran working sessions with each engineering team to explain design intent and resolve implementation questions before they became blockers.

Knowledge transfer
The spreadsheets, Asana structure, and working sessions were not just about closing findings. They were about leaving each team capable of working independently so the next round would not start from zero.

Key Design Decisions

Annotation as communication layer

Accessibility remediation fails most often in the gap between what a designer intends and what a developer implements. Every finding annotated in Figma, tied to a spreadsheet with severity, WCAG criterion, and Figma link. Developers had everything in one place without a sync for every item.

Figma files organized by screen, not by finding type

The standard approach groups findings by category: all contrast issues together, all focus order issues together. I organized by platform screen instead. A developer working on the login page opens one Figma page and sees every issue for that screen. The organization matched how development work actually happens, not how audits are reported.

Finding triage by responsible discipline

Every finding assigned a discipline: design, development, design system, or tech writing. Each team filtered to their own work without wading through findings that were not theirs. Progress tracked per discipline, blockers flagged to the right person without a group sync.

Solving the design problem underneath the finding

A finding on the Schedule Overview Weekly View flagged that activating "Expand All" did not move keyboard focus to "Collapse All," a WCAG 2.4.3 Focus Order failure. The standard remediation recommendation was JavaScript focus() management. That would have closed the finding.

I made the case for a different solution instead. Two buttons for a binary expand and collapse action is the wrong component. It is a toggle. Replacing both with a single toggle from the design system solved the focus problem structurally: the same element stays focused when its state changes, so focus management becomes unnecessary, and keyboard tab stops are reduced. The toggle pattern already existed in the platform as Condensed View, so the solution was consistent with existing usage.

The finding was closed by using the correct component, not by patching the code.

Two WCAG failures. One component swap and one copy edit.

Login screen, Workforce Management Mobile. WCAG 1.4.4: helper text cut off at 200% font size. WCAG 3.3.2: field relied on placeholder text as its only label, which disappears on input. Standard fix: two separate patches. Instead, I replaced the field with the Alloy design system's text field component, which has a persistent visible label by default, and moved the helper text inside the field as shorter placeholder text. Both findings resolved in one decision.

Component-level fixes over screen-level fixes

Where a finding pointed to a systemic design system issue, I escalated beyond a single screen. Updated contrast tokens, focus order logic, accessible labels, and reflow behavior at the component level for high magnification scenarios. A fix at the component level compounds across every surface that uses it. Full-screen reflow on the legacy iframe-based platforms was a separate constraint, addressed in trade-offs.

Trade-Offs

Delegating low-risk fixes to engineering
Some low-risk findings were handled by developers with written guidance rather than a full Figma redline where schedule pressure was high. That required precise documentation and trust in the engineering teams.

The reflow limitation
Reflow compliance (WCAG 1.4.10) could not be remediated across the legacy platforms, which were built with iframes. Achieving it would have required a full frontend rewrite, outside the scope of a remediation engagement. I applied it where achievable, such as certain modals, and had to make the case to each product team that full screens were not fixable within the existing architecture without that rewrite. This was documented as a platform-level limitation rather than patched around, the same constraint that later justified the full rewrite approach taken in the spend management platform.

The shifts calendar
The shifts calendar in the Workforce Management Platform relied on colored circles to distinguish shift types, a WCAG 1.4.1 failure. The correct fix required a full calendar redesign beyond the remediation scope. The finding was documented and handed to the product team as design debt. Applying a partial fix would have created visual inconsistency worse than the existing state. Knowing when not to fix something is part of the work.

Outcomes

• 1,274 confirmed findings resolved across six platforms

• All six platforms achieved a 90+ accessibility score as measured by Level Access

• All six platforms brought into WCAG 2.2 and Section 508 compliance, serving 200,000+ users

• Concentrix Game Changer of the Month December 2025

• Recognized at the Argentinian Townhall 2025

• Repeatable workflow and documentation established across all six teams for independent use

Reflection

What this project made clear is the difference between fixing accessibility and building for it. The findings were the starting point, not the goal. The goal was to leave each team with patterns they could reuse and a process that did not depend on me being in every room. Getting there across six teams with different cultures and familiarity with accessibility was the real design problem.

This work was done entirely manually before AI tooling could meaningfully assist. That direct experience is what makes it possible to know exactly where automation would have the most impact, which is the basis for a separate project on AI-assisted accessibility auditing and remediation.