Please use the boxlink for a recorded demo of our project.

We could not upload our video to youtube or the other proposed platforms but have included a boxlink to a recorded demo: https://ibm.box.com/s/v4kljjav4qi7ls2c4igdd8nry0nwax6i

Introduction

The IBM Ireland Innovation Exchange been working with health care policy makers for a number of years to develop methods and tools for collecting and sharing health related information. One specific requirement of the policy makers is to engage the public in conversation and gather sentiment regarding key policy changes and reform. Traditionally these are done through paper based surveying, or, for a handful of major reforms via citizen assemblies and workshops. These manual approaches don't scale and are not representative of wider society.

Context and Challenges

What about using bots? A significant limitation of many bots is that they are traditionally used to provide answers to predefined questions. This can be quite effective for insurance, retail or even technical support helplines where a significant amount of queries and therefore answers are the same. Their application to health based industry is typically suitable only to basic use cases e.g. "For more information on X please see Y", “If you are doing A please do B". Modern implementations of bots can even match new variations on questions with existing questions in their database, but this still provides little flexibility and no conversational capability. After all, we normally only understand an individual’s story by engaging in conversation, asking them questions and probing their answers. In short yes/no questions supported with answers based bots are not sufficient for this unmet need.

The Social Campaign Manager

The Innovation Exchange have implemented a prototype of the IBM Social Campaign Manager as a tool to solve this unmet need. The application allows completely non-technical users to develop questionnaires using Watson Assistant that include yes/no, multiple choice and most importantly free form questions. After the campaign owner has structured the conversational flow of the questions they want to ask they can publish the campaign to Twitter. To "Publish" in this context means a tweet is sent out via Twitter as an invitation to engage in a conversation. For example the health service or NGO could send a Tweet - "Do you want to help slow the virus and #FlattenTheCurve? If so let’s talk!". A participant in the campaign can then engage in a private conversation within Twitter or a generic chat window on the health services website. The campaign structures the conversation using Yes/No Questions e.g. " Are you self-isolating?", it can measure education levels on specific topics e.g. "What is social distancing is it ...A..B....or C". The key value add can also come from a free form questions in response to how the participant is answering binary questions e.g. "You have indicated that you are self-isolating, can you please explain what motivated you to do that?", from here the user can tell their own story.

As the parallel conversations are managed by Watson Assistant information is being gathered and added to a dashboard presented to the campaign owner e.g. the campaign owner. In the dashboard demographics including age, gender and location of the participants can be shown. Graphs and pie charts are used to represent statistical information, word clouds and sentiment analysis tools are used to try to understand the meaning behind participants answers. The dashboard will also show if someone gave an answer to one of the questions that was not expected e.g. to a multiple choice question allowing the campaign owner to identify if participants are understanding the questions or if editing is required. At any time the owner can amend and republish the campaign.

Conclusion

Multiple campaigns can be written, launched and managed in parallel e.g. one could be focused on gather information around the virus, the other could be aiming to identify cohorts feeling most stressed or isolated and provide them with education to help them. In the Covid use case, the first campaign could initially focus on the employees within the health service, carer organizations or any frontline group rather than the general public giving them a voice, capturing their concerns and taking actions available to help with them.

Acknowledgements

This tool was developed in collaberation with the MIDAS project which was funded by the EU Horizon 2020 framework with Grant ID 727721 https://cordis.europa.eu/project/id/727721

Please use the boxlink for a recorded demo of our project.

We could not upload our video to youtube or the other proposed platforms but have included a boxlink to a recorded demo.

Share this project:

Updates

posted an update

The direction of the tool has been evolving as we went along. After a successful introduction of a chat interface through Twitter, we introduced a secondary input for people who may not be as familiar with the social network or do not wish to set up an account to talk to participate in the campaign. Adding a separate web client for a direct chat system into the Social Campaign Manager has been a breeze due to the modular way it is build with.

Log in or sign up for Devpost to join the conversation.

posted an update

We had a lot of the code for this tool developed as part of a long standing collaberation with Healthcare policy makers and R&D stakeholders. We have been working to repurpose and align with use cases that could enable the broader community to developer AI enable applications to have a positive impact in these challenging times.

Log in or sign up for Devpost to join the conversation.