Benefits portal extraction
Playwright-based extraction across all PlanYear plan types, handling Ant Design quirks and outputting clean structured datasets.
// OUTCOME
- 01Replaced manual plan-by-plan data entry for benefits consultants.
- 02Reliable extraction across plan types and edge cases.
- 03Clean structured output ready for downstream use.
// PROBLEM
Benefits consultants needed structured data from the PlanYear portal across multiple plan types (medical, dental, vision, life, etc.). The portal uses an Ant Design UI with dynamic component quirks, pagination, and conditional fields. Manual extraction meant clicking through every plan, every section, copying data into spreadsheets — slow, error-prone, expensive.
// SOLUTION
- 01Logs in programmatically and maintains session state.
- 02Navigates across all plan types — medical, dental, vision, life, and more.
- 03Handles Ant Design's dynamic component quirks, pagination, and conditional fields.
- 04Scrapes structured plan data reliably even when the UI shifts.
- 05Outputs clean normalized datasets ready for downstream consulting workflows.
// STACK
PythonPlaywrightAnt Design DOM handlingSession/auth managementData normalization
WEB AUTOMATIONDATA EXTRACTIONPLAYWRIGHTBENEFITS