First, we wanted to go for the Sanofi challenge because we love Amazon. However, on Saturday morning started discussions with the Philips challenge owner Shari and Lena Hannig, who is a midwife in Berlin, and realised that we had already many ideas in this field. Both of us are already drawn into project connected to female health, that is our passion. So we decided not to fight nature and develop the solution that we had in mind, AskMidwife on Facebook Messenger. While talking with Shari and Lena, we identified the most important pain points of midwives, and build a chatbot solution for that. This would be the natural extension of our already working chatbot, Izzy.
What it does
AskMidwife has 3 main features referring to the 3 biggest pains of midwives. 1) Answering some basic questions on pregnancy and breastfeeding, saving time for midwives that they would spend on answering unnecessary calls and messages 2) Diary function where women log how they feel and what's challenging during pregnancy. This is free text input that is connected to IBM Watson's personality analysing feature, giving and overview of the patients mental status. 3) Appointment scheduling making it easier and faster for both midwives and patients to manage their calendar.
How we built it
The software is built in python, with Messenger and IBM Watson integration. One big part of the project was to identify the exact needs of midwives and to design the conversation flows. Then we had to figure out how to support these needs with the relevant, new features (diary, appointment scheduling). We used IBM Watson to deliver some personality analysis on free text input.
Challenges we ran into
Deployment from AWS, getting conversation flows running correctly being able to handle every possible situation.
Accomplishments that we're proud of
Developing new features that analyses free text input, not using only keywords. Developing completely new content based on the needs and inputs of midwives.
What we learned
Needs of midwives and how automation could help them, how IBM Watson should/could be used in real products.
What's next for AskMidwife
We will put this project further by adding the content that we developed to our current product, and provide a better service to our pregnant users. To do so we will establish a collaboration with midwives to work on the content together, and also to make live the appointment scheduling feature with calendar integrations.
Log in or sign up for Devpost to join the conversation.