BEBO is needed because during the spring of 2020, approximately tens of millions of Americans lost their jobs as a result of the COVID-19 pandemic. According to the New York Times, the unemployment rate as of mid-May reached 14.7%, generating comparisons to the Great Depression. State unemployment agencies were not prepared to handle such a large number of applicants all needing service at once. Workers seeking benefits, or even just answers to simple questions, often could not get through on the phone and were confused by the application system, eligibility requirements, and pandemic-specific programs. Sadly, millions who were eligible did not even bother applying because the system was so hard to navigate. The inability to secure benefits can have dire financial consequences that impacts entire families. Of all those impacted, women suffered the greatest losses and were hit the hardest.
BEBO is unique in that she was built on extensive user research done by our team. We investigated the pain points of people using the system by going through several social media posts and came up with more than fifty questions to address them. BEBO is built to focus on providing personalized answers to people in need of help, rather than providing a generic FAQ-style question-answer.
We used Dialogflow framework to design our chatbot. We designed the conversations such that each dialogue retains a contextual memory to pass on to the next conversation. We also tested and assigned lifespans to each contextual memory, which may vary depending on the context. This helps BEBO narrow down to obtain the right responses to the user questions. We also tested BEBO extensively so that any variations of the user question or phrase are picked to search for the right response. We also took special care to choose words so that BEBO can sound human.
If you have any questions or suggestions, please email us at whowteam@gmail.com