AI Assisted Testing for Injury Management Platform

We developed an AI-assisted test automation program for an enterprise platform designed to manage personal-injury cases. The platform connects medical providers, law firms, and internal staff, supporting key processes such as medical record management, appointment requests, medical billing, and plaintiff funding. The platform comprises a core application, a modernized legacy case-management system, and a legacy appointment-tracking API, requiring test automation across multiple interconnected components.

In its earlier state, every release depended heavily on manual regression: 2.5 days to 2 weeks per cycle, 2–4 rounds per release, and 6–8 QA engineers plus developers getting pulled in to “swarm.” In a certain instance, one of the sprints delivered only 44% of its committed story points because regression consumed QA bandwidth, and a scheduled release slipped a full cycle after regression found blocking bugs. In order to address these cases, the QA team introduced an AI-assisted testing system: specialized agents for test authoring, live-fixture discovery, pull-request review, and regression reporting — each guided by written repository steering documents and backed by deterministic Node.js scripts, so that the AI agents are never left to guess.

The agents now assist engineers in drafting and extending Playwright test specs, resolving live test data per environment tier, reviewing every pull request against a fixed checklist, and narrating root causes in sprint regression reports. All this runs inside one shared automation repository covering every product, with results reported automatically to Microsoft Teams.

This AI-assisted workflow helped grow the automated test suite by 174% over 12 months (2.7×), while a full regression of 2,187 tests now completes in under 30 minutes at a 95.6% pass rate — work that earlier took days and multiple manual rounds.

Client Name: Confidential
Industry: Healthcare
Location: USA

QA Automation & AI Enablement

TypeScript, Playwright Test, Claude Code, steering documents; Bitbucket Pipelines, Azure DevOps, Azure Playwright Service; Jira, Xray, Confluence; C# / .NET 8 (xUnit, NUnit, Moq, FluentAssertions, NetArchTest); Angular, React, Jest, React Testing Library, MSW; Page Object Model; Microsoft SQL Server.