Skip to content

Ai4impact co-op | Fall 2023- Spring 2024 project

A Healthier Democracy AI

Delivers over $4M in federal aid by helping low-income families apply for benefits directly in healthcare settings via an AI assistant.

ABOUT THE PROJECT

Helping low-income residents apply for unclaimed Federal Aid benefits in a healthcare setting. This AI assistant helps low income families apply for benefits in the emergency room.

Impact: Over $4 million in benefits disbursed. One of the students was hired to be an intern with EOTSS following her co-op.

Partner: Link Health

Status: In use.

Introduction to ScreenWise

The ScreenWise initiative addresses the challenge of centralized eligibility screening for federal benefit programs. Current systems require individuals to navigate fragmented program requirements, often leading to confusion and underutilization. ScreenWise leverages AI to streamline the process, making it accessible and efficient for non-profits and individuals seeking multi-program benefit eligibility.

The Gap between Eligibility and Enrollment

Navigating federal benefit program eligibility is time-intensive and complex. People struggle to sift through varying requirements across programs like SNAP, WIC, and Lifeline. Not knowing their eligibility status was the top reason among 63% of Massachusetts residents who were eligible but not enrolled in SNAP in a survey by the Boston Food Bank. Complexities like these leave billions in unclaimed benefits annually, over 80 billion nationwide. By centralizing and automating this process, ScreenWise eliminates barriers, allowing organizations to screen for program eligibility efficiently.

Overview of ScreenWise

ScreenWise simplifies federal benefit eligibility screening by combining two tools:

  1. Generative AI Eligibility Extractor: Upload program documentation, and the tool outputs a structured JSON file detailing eligibility thresholds, criteria, and screening questions.
  • Purpose: Transforms federal program documentation into structured eligibility data.
  • Output: A JSON file encoding programs, questions, and criteria.

2. Eligibility Screener Builder: A dynamic logic engine that tracks user responses, refines questions, and provides a list of eligible programs along with savings estimates and application links.

  • Purpose: Provides a web-based, interactive tool for screening multi-program eligibility.
  • Features:
    • Dynamic, real-time logic adjusting based on user responses.
    • Conditional eligibility handling (e.g., qualifying via income thresholds or regional criteria).
    • A React-based interface embedded into non-profit websites for seamless integration.

These solutions empower non-profits and individuals with a secure, scalable platform for discovering benefits while maintaining data privacy.

Informed by User Research

ScreenWise is informed by research conducted by the Burnes Center, revealing that fragmented eligibility systems and lack of understanding are primary barriers to benefit access.  Referencing conversations with experts and people applying for benefits ensures the tool aligns with real-world needs, offering practical, user-centric solutions.

The tool is functional and has been tested internally by AI for Impact team members. The tool has not yet been tested by organizations or tested for public release.

Introduction to ScreenWise

The ScreenWise initiative addresses the challenge of centralized eligibility screening for federal benefit programs. Current systems require individuals to navigate fragmented program requirements, often leading to confusion and underutilization. ScreenWise leverages AI to streamline the process, making it accessible and efficient for non-profits and individuals seeking multi-program benefit eligibility.

The Problem with Federal Benefit Eligibility Screening

Navigating federal benefit program eligibility is time-intensive and complex. People struggle to sift through varying requirements across programs like SNAP, WIC, and Lifeline. This complexity prevents 63% of potential applicants from pursuing federal aid, leaving billions in unclaimed benefits annually. By centralizing and automating this process, ScreenWise eliminates barriers, allowing communities to discover their eligibility efficiently.

Overview of ScreenWise

ScreenWise simplifies federal benefit eligibility screening by combining two AI-powered tools:

  1. Generative AI Eligibility Extractor: Upload program documentation, and the tool outputs a structured JSON file detailing eligibility thresholds, criteria, and screening questions.
  2. Eligibility Screener Builder: A dynamic logic engine that tracks user responses, refines questions, and provides a list of eligible programs along with savings estimates and application links.

These solutions empower non-profits and individuals with a secure, scalable platform for discovering benefits while maintaining data privacy.

Informed by User Research

The ScreenWise design is informed by research conducted by the Burnes Center, revealing that fragmented eligibility systems and lack of understanding are primary barriers to benefit access. Collaboration with stakeholders like the Boston Food Bank ensures the tool aligns with real-world needs, offering practical, user-centric solutions.

Current Work: ScreenWise’s Features

1. Generative AI Eligibility Extractor

  • Purpose: Transforms federal program documentation into structured eligibility data.
  • Output: A JSON file encoding programs, questions, and criteria.

2. Eligibility Screener Builder

  • Purpose: Provides a web-based, interactive tool for screening multi-program eligibility.
  • Features:
    • Dynamic, real-time logic adjusting based on user responses.
    • Conditional eligibility handling (e.g., qualifying via income thresholds or regional criteria).
    • A React-based interface embedded into non-profit websites for seamless integration.

Future Work

Next Steps

  • Testing ScreenWise with non-profit partners will refine its functionalities and scalability. Future updates will focus on integrating state-level program data, optimizing performance, and expanding accessibility.

PROJECT TEAM

  • Sarah Klute
  • Rishabh Saxena

We use cookies to improve your experience on our sites. By continuing to use our sites, you agree to our Privacy Statement.