The Employee Benefits Security Administration
The Employee Benefits Security Administration commits to educating and assisting more than 153 million workers, retirees, and their families covered by the benefit plans they oversee. This includes overseeing more than 765,000 private retirement plans, 2.8 million health plans, and 619,000 other welfare benefit plans, which collectively hold about $13.4 trillion in assets.
To ease inter-agency data constraints that led to system redesigns impacting EBSA in terms of schedule and scope, WebFirst was able to partner with EBSA to apply Lean-Agile principles to reduce the strain on EBSA's team by focusing on resourcing and risk management while continuing to deliver high-quality deliverables.
In support of the Employee Benefits Security Administration's (EBSA) mission to ensure the security of retirement benefits for America's workers, WebFirst developed the Retirement Lost & Found Database with advanced machine learning (ML) and AI capabilities within a modern Drupal architecture. WebFirst leveraged the Department of Labor's (DOL) existing Drupal application stack and cloud services to create a portal that enables individuals to locate missing retirement benefits and allows plan administrators to manage plan data.
Our customer experience (CX) process included conducting targeted user interviews to inform mockups, prototypes, and UX artifacts. The architecture consisted of decoupled Drupal with React, chosen for its standardized API query handling and strong Drupal module support. For security, WebFirst integrated multi-factor Login.gov authentication for the public-facing portal, with automated role-mapping for secure access, and implemented Azure single sign-on for system administrators.
Moreover, WebFirst built an ETL and Machine Learning (ML) text extraction system to process approximately 3 million Form 8950 records using AWS Glue and Textract - which is capable of parsing data from diverse document formats such as PDFs and PNG files at a low computing cost. WebFirst’s experience building ML/AI tools was essential for processing the extensive collection of unstructured data and ingesting encrypted data in CSV and XML formats.
Our project management team oversaw this effort by employing Agile methodologies (Scrum with Jira), including the Section 508 Trusted Tester process, to enhance site usability and minimize errors. The strategy included employing USWDS 3 components - which adhered to Section 508 and WCAG 2.2 AA standards - while maintaining EBSA’s brand guidelines, to arrive at a consistent, expected, and successful resolution.
