IU Luddy School Students Create AI Chatbots To Aid USDA SNAP Program



On Feb. 26 –March 1, MetroStar partnered with graduate students from the Indiana University (IU) Luddy School of Informatics, Computing, and Engineering HCI/d (Human-Computer Interaction and Design) program to showcase their skills and how they can make an impact in a critical area during a weekend design challenge. Our challenge centered around COVID-19’s impact on USDA’s SNAP (Supplemental Nutritional Assistance Program) program and how design and emerging technology could be leveraged to support the at-risk populations that rely on this vital program.

Chatbot Challenge Background

The coronavirus has impacted everyone across the world in ways we may not feel for generations. Our Government has created numerous programs to aid fellow Americans through times of struggle and help them get “back on their feet.” One of those programs is USDA’s Supplemental Nutrition Assistance Program (SNAP) who’s purpose is “Ending Hunger and Moving People Forward to Self Sufficiency.”

The Federal Government pays the full cost of SNAP benefits but shares the costs of administering the program with the states. States have the flexibility to streamline program administration, remove barriers to enrollment, and adopt policies that will ensure benefits are reaching people in need. While it’s beneficial for states to be responsive to the needs of their communities and residents, it can provide a challenge for potential applicants trying to find the official information they seek.

Within each state, SNAP is managed by varying agencies; those agencies have their own priority programs taking away the finite digital expertise, available technology, and limited funding necessary to craft, deploy, and successfully manage a program of this size and reach. Each of these factors has the potential to negatively impact the program’s effectiveness and perception. MetroStar challenged the students to think about ways to improve a user’s experience through the integration of centralized chatbot framework each state could leverage and tailor to their specific implementation along with overcoming the usability pitfalls commonly associated with chatbots.

Chatbots are a branch of Artificial Intelligence (AI) that takes human language input through voice commands, text chats, or both and understands what is being asked of it. The AI then provides a relevant, correct answer based on the knowledge it has.  

The value for implementing a national-level chatbot framework with capabilities that can be tailored at the state level is the incorporation of mature human-centered design (HCD) methodologies that guide the rationale and decision-making for the technology and how it should be implemented and evolvedThrough an HCD-first approach, we are able to better understand the needs, frustrations, and barriers to adoption of those that:

2) have been affected by COVID-19 (directly or indirectly) and now need apply for SNAP benefitsand

3) are responsible for administering SNAP at the state level.

A human-centered experience would help to improve the findability of information for potential applicants and decrease their frustrations spent trying to navigate disconnected systems and decipher confusing informationA better user experience that aids decision-making and understanding of eligibility requirements would also decrease the administration time and costs spent reviewing ineligible applications and help desk calls. 

This interconnected framework would improve the chatbot’s responses improving decision making and connecting SNAP users with the correct state-specific information they need in a manner that is understandable, supportive, relevant, timely, and accessible.

The Design Challenge

On Friday morning during the event kick-off, IU Luddy students were divided into teams of three and tasked with taking the weekend to develop a chatbot framework that could be reused for each state’s individual application process, which would establish a seamless and cohesive user experience for SNAP applicants across the country.

On Monday morning, the teams gave 15-minute presentations on their approaches to a panel of judges from MetroStar’s Client Solutions Group (CSG) and USDA UX teams. Snapberry, Snapdroid, and Ezra are the names of the three chatbots that were developed to help SNAP applicants get their questions answered and applications filled out.

“The MetroStar team was extremely impressed by the thoughtfulness and quality of the work the students developed in such a short timeframe,” Jason Stoner, MetroStar’s Director of Experience Strategy, said.

Winners of the challenge are being presented with the financial resources to purchase software to further the development of their UX expertise. Winners also have the option to interview for a UX internship at MetroStar and the opportunity to present their work to the entire MetroStar company.

“The challenge was a valuable and rewarding experience. It gave us a chance to apply our developing skills in a design space with strong social merit,” said Alex Dlugosz, an IU Luddy student and member of the winning team (Ezra).

“MetroStar was thrilled to work with the HCI/d graduate students from the IU Luddy School and looks forward to partnering with IU again on further challenges and student projects,” Anne McCombe, MetroStar’s Business Relationship Manager, said. “We are always passionate about bringing innovation to the government through partnerships between industry and academia.”