COVID-19 Chatbot


A COVID-19 screening chatbot tool 

Role: User Researcher, Market Researcher 


What is a Chatbot?

A chatbot is a computer system that allows users and a software application to communicate with each other, this can be using spoken or written language. Chatbots interact with users basing their responses on natural language dialogues. Chatbots are beyond a set of commands, and it’s claimed that the use of chatbots is trying to imitate a similar way to a human-to-human interaction asking for a transaction. This is why it’s essential to approach a chatbot design using a user-centered method to best accommodate users’ needs.


Design objective

To condense and standardize the access to information related to COVID-19. The purpose is to take a complicated set of recommendations that outline how and when to access care and distill it into a set of simple questions. Based on the answers, provide personalized recommendations to access care and maintain people safe.


Design Process

During the research phase of this project, I realized that most of the COVID-19 related websites offered three main pieces of information for patients: COVID-19 screening, finding testing locations, and frequently asked questions. Since users seek quick, accurate responses when searching for information, a chatbot can provide this without the user having to do a web search.

COVID-19 screening chatbot

This chatbot can help users find testing locations near them, ask general questions about the virus, do a quick COVID-19 screening that tells you what you should do; get tested, stay isolated, or if your risk of contagion is low. The information used to create this chatbot was gathered through indications of the Center for Disease Control and from numerous global health organizations. While each person has individual health conditions, many of the COVID-19 screening processes can be standardized to provide up-to-date guidelines and information.

Throughout the world there are increasing cases of COVID-19 and access to resources and information can be difficult in developing countries. I decided to create a Chatbot to be implemented on Facebook Messenger because I found that in developing countries is common that people have access to smartphones and social media but not access to web browsing. This is mostly because some carriers in these countries offer very limited data, but offer unlimited access to social media such as Facebook, WhatsApp, and Twitter at a low cost. Due to this, for some people, the only access they have to online communication is through Facebook where oftentimes misinformation is spread at a very high rate.

Competitive Analysis 

From this comparison, the Chatbot from Providence was the one that gave the most enjoyable, quick responses. This chatbot was able to interact fast and give accurate responses and almost felt human. To start the chatbot was confusing because there were a few legal questions you had to answer before even starting it. The intent for my chatbot is to translate a similar feeling but for Facebook Messenger.


Information architecture 



Onboarding experience 

The initial interaction provides three different choices for users to interact with the bot. The options were chosen based on research trends when users are looking for information on COVID-19. 

Answering general questions

The database will have information on common questions.  

Finding locations 

The chatbot is able to find testing locations close to a user’s home. 

Project retrospective

The most challenging about this chatbot is making sure to empathize and synthesize information in less text. I think the biggest challenge is recompiling information and making it accessible to everyone while avoiding the chatbot getting too technical and robotic. This process goes from gathering information from official sources and rewriting information. Another concern I had, it’s that the messages are too long and that users might not want to read all of the information, but I think I was able to synthesize the information and find a balance on the type of information shared.

I’m the proudest of the final result, as I think it works well and it’s able to provide relevant information fast. I enjoyed coming up with a problem and learning to use Botsociety. I think that bots can be an effective way to come up with solutions for different markets.