As more vaccines for COVID-19 gain approval for public use and production increases, there’s a new barrier to reaching herd immunity – vaccine hesitancy. Concerns about vaccine safety and misinformation contribute to a large portion of the population expressing reluctance to get vaccinated, enabling the further spread of the disease.
Here’s the good news – this is a challenge we can meet.
Digital communications can help address people’s concerns and deliver fact-based information at scale. Specifically, information delivered through chatbots by trusted organizations positively impacts an individual’s attitude toward the COVID-19 vaccine.
The positive influence of bots is twofold. First, chatbots allow people who have concerns to ask their questions in a private, non-judgmental space. Second, chatbots are conversational, and mimic small-group discussions with trusted healthcare providers that are more effective at influencing perspectives than one-way arguments.
To empower more organizations to quickly create chatbots that address vaccine hesitancy, today we’re launching a Twilio Quick Deploy app for vaccine FAQ chatbots. Organizations working on the front lines of vaccination campaigns can deploy the template in minutes with zero coding required.
A flexible, intelligent FAQ bot for COVID-19 vaccines
This Quick Deploy app provides a framework for creating a conversational messaging bot that is completely customizable for any public, private, or nonprofit organization’s needs. With just one click, you can deploy the app in your Twilio account and begin testing a proof of concept of a vaccine chatbot using one of two pre-built agents. Or, you can use the included tooling to create your own conversational bot from a vaccine information source of your choosing.
This generalizable template enables you to skip the grunt work and get right to customizing a bot to serve your audience’s needs. After all, time is precious.
There are three key pieces to this Quick Deploy app.
- A pre-built Twilio Studio Flow that orchestrates the conversation based on user inputs over SMS. The Flow can be extended to any of Twilio’s supported messaging channels, including WhatsApp, and chat
- An integration with Google Dialogflow that handles the natural language processing to analyze user intent and serve up relevant responses to inbound queries
- Two chatbot agent profiles that are pre-trained on information from trusted health sources to manage inquiries:
- A chatbot for U.S. residents based on an FAQ from the American Medical Association
- A chatbot for Indian residents based on an FAQ from the Ministry of Health and Family Welfare
Here’s what it looks like in practice:
End users will be able to interact with the bot via SMS and ask it common questions like: "Are the COVID-19 vaccines safe?," "Will the vaccine give you COVID?" and "Can COVID-19 vaccines get you sick?" Organizations using the Quick Deploy app can extend or adapt the chatbot by continuing to train the Dialogflow agent. For instance, you may want to add new question and answer pairs, or tailor the response to an existing question to be more relevant to the audience you are trying to reach.
The Quick Deploy app is part of the CodeExchange, a library of open-source Twilio templates. Through the CodeExchange interface, you will be able to deploy the app with zero coding required, simply by following the pre-configuration steps and filling out the form to wire up the app to your Twilio project and phone number, as well as authenticating properly with your Google Cloud account:
Included in the Quick Deploy configuration is a language code environment variable that allows organizations serving global audiences to deploy chatbots in any of the 55 languages supported by Google Dialogflow.
The app comes prepackaged with an included Studio Flow that will orchestrate the conversational chatbot interface. This will route SMS messages sent to your configured phone number, forward the contents of the user input to Dialogflow for processing, and respond back based on the detected intent. While the Studio Flow works "out-of-the-box" with no extra development required, it is entirely customizable to the needs of your organization:
Extending the bot to serve more communities
The two pre-built agents included in the Quick Deploy app are just the starting point. Vaccine hesitancy is a global problem, and serving diverse communities globally will require adding more agents trained with local information, language, and cultural context. For instance, a chatbot for addressing vaccine hesitancy among rural Americans is very different from information needed for Rohingya refugees in Bangladesh.
We invite anyone who is interested in serving their local community to build and train additional agent profiles and contribute them to the open source community project.
To make this easier, we’ve included some helper tooling to build new chatbot agents from scratch. You can use this included script to extract question and answer pairs from published vaccine FAQs, and automatically train a Google Dialogflow agent to respond appropriately to user inquiries based on your chosen data source. The script will even train the Dialogflow agent on paraphrases so it will respond to similar versions of the question asked by a user.
Together, this project can grow into a global repository of vaccine hesitancy chatbots that can help everyone get the information they need to feel confident about receiving a COVID-19 vaccine. We hope you’ll join us in this initiative, and we can’t wait to see what you build.
Get started with the COVID-19 vaccine hesitancy Quick Deploy app here.
Twilio supports organizations working to end the pandemic
To support global vaccine distribution and administration, Twilio.org is providing $1 million in product credits for organizations that are helping underserved communities get vaccinated.
To learn more about funding opportunities and apply for credit, visit Twilio.org today.
Big thanks to Sudheer Chekka, Ishita Aggarwal, Aleksei Kuryla, and Ci-Ci Thompson for their contributions to creating this application.