EverDriven
Accelerating Software Quality Improvements with Claude
Baufest built Claude Code into the Playwright test framework for TripHQ, EverDriven's trip-management platform. Automating a new test case now takes about 2 hours instead of 2 days, and checking for visual defects takes about 5 minutes instead of 2 hours. Test coverage also expanded into security, accessibility, and internationalization.
- Industry
- Capabilities
- Partners
- Regions
- 🇺🇸 United States
- Published
- October 2026
The Company
EverDriven provides student transportation in 38 states and hundreds of school districts. It coordinates about 10,000 trips a day, many for students with special needs. Its dispatch, field, and routing teams run those operations in TripHQ, so the platform has to be reliable.
The Challenge
TripHQ is the portal EverDriven's internal teams, transportation providers, and school districts use to manage trips. In 2026 it was moving from MVP to production, adding real-time trip tracking, role-based permissions, data isolation between providers and districts, and Spanish-language support.
Every feature had to meet EverDriven's quality bar: at least 80% automated test coverage, a 95% nightly end-to-end pass rate, and no Sev-1 or Sev-2 defects at release. With each new test case taking about two days to automate, testing was falling behind development.
The Solution
Baufest's Quality Engineering team, embedded in EverDriven's sprints, extended TripHQ's existing Playwright framework with Claude Code. They connected it to the same three sources a test engineer relies on:
- Jira, for the requirements behind each story
- Playwright (via MCP), to navigate the running application
- Figma, to compare screens against their designs
Custom skills and commands encode the project's conventions, so the tests Claude Code generates fit the framework's structure and don't need rework. A QE specialist reviews every test before it's merged.
- 2 hours
- To automate a new test case, down from 2 days
- 5 min
- To detect visual defects, down from 2 hours
- +14%
- Security test coverage
- +18%
- Unit test coverage
- +20%
- Coverage across existing functional, visual, accessibility, and internationalization tests