The latest from DXC’s Applied AI Center of Excellence: A chatbot to help remote workers de-stress

remote work

Lee, an online personal counselor, helps employees who struggle with feelings of mild anxiety or loneliness while working from home. Rather than cope alone, they talk with Lee to destress.

“You’re lonely?” Lee might say. “I’m sorry about that. Can you tell me more about how you’re feeling?”

Those words might not seem surprising from a human. But Lee is an AI-powered cognitive chatbot, and the latest creation of the Artificial Intelligence Center of Excellence (AI CoE).

AI moonshot

We developed Lee as part of what we in the AI Center of Excellence (CoE) call an “AI moonshot.” Moonshots are innovative and ambitious pursuits that aim to help DXC customers solve real-life business problems through the application of artificial intelligence. Graduate students, who are also DXC professionals, work to demonstrate what AI can do, with a focus on practical applications that can be integrated into existing enterprise systems.

We focus on moonshots because many of our customers lack the time or energy to explore innovations that could improve their business in the future. Others have a hard time imagining how new and different technologies could help them. By incubating ideas and bringing them to life in the form of prototypes, the AI CoE helps customers not just imagine the potential, but also see it.

Helping remotely

The cognitive chatbot project that developed Lee began this past April. We realized that some remote employees, many of whom were working at home for the first time as a result of recent physical distancing and stay-at-home guidelines, might be suffering from anxiety and mild stress. To help, we came up with the idea of creating a mental-health counseling chatbot.

“It is not always feasible for people to visit a therapist,” explains Hemanth Ariktala, an AI CoE Information Engineer and the chatbot project’s product owner. “Our challenge was to make mental health support much more accessible with the help of a digital human bot.”

To create the bot, a small team from the AI CoE worked with Dr. George Mathew, DXC’s former chief medical officer, and two psychologists at Drexel University to create a series of chatbot scripts.

One of the academics was Dr. Adrienne Juarascio, an assistant professor of psychology at Drexel, who has studied the use of digital technology to provide just-in-time interventions for people suffering from depression, anxiety and eating disorders. “Getting resources to as many people as we can during this pandemic is a top priority,” she says.

To do this the team mapped the AI component of their chatbot to these intents. They also drew on research showing that cognitive behavioral therapy (CBT) delivered over the internet can be as effective as therapy delivered face-to-face, especially for relatively mild problems such as anxiety and stress.

Lee will see you now

To deliver CBT online, our team created a UneeQ digital human avatar — an animated woman named Lee, who uses Google DialogFlow to “understand” the user’s spoken questions and deliver a verbal response.

To keep the conversation flowing, Lee is also equipped with conversational small talk, such as “Thanks for sharing,” “Ah, okay” and “Got it.” When appropriate, she can also share recommendations and links to helpful resources via pop-up text boxes.

Cognitive chatbot

Lee, a cognitive chatbot, recommends a helpful link.

Throughout the project, our team employed Agile development methods. These included daily scrums, which are brief team sessions to keep everyone caught up, and regular meetings with our medical experts. We brought Lee to life using two-week development sprints, which allowed team members to incorporate ongoing feedback into the overall design.

Working with Agile methods was especially intriguing for Jina Huh-Yoo, an assistant professor at Drexel’s College of Computing & Informatics who worked with our chatbot team. “In academia, we don’t get to do this type of rapid Agile project, constantly iterating and improving,” she says. “This was a great opportunity for me to expand my areas of interest and collaborate with people in industry.”

Lee was inspired by what we were experiencing with the pandemic; however, supporting mental-health and remote work isn’t limited to our current situation. Mental health services, whether in-person or made more widely available by an AI, is an ongoing need.  And it is a need that can be met swiftly by AI. We completed our entire chatbot project in just 12 weeks. As we tackle other moonshot projects with customers in our AI CoE, watch this space to learn about the transformative power of AI.


Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.