Providing Hope in the Form of a Chatbot

At AdventHealth, we believe health should be measured in terms of the whole person — body, mind, and spirit. And it’s our promise to help our patients feel whole through compassionate care and world-class expertise.

Our health system, like many others, was put to the test at the start of the pandemic as we had to adjust the way we operated in a short time span. As states were put under stay-at-home orders, most of our hospital campuses and hundreds of doctor’s offices, urgent care clinics, imaging centers, and labs turned a great deal of their attention to servicing COVID-19 and emergency patients.

We needed to find digital tools to provide the same level of whole-person care and connection.

At the start of the pandemic, our patients understandably had many questions and concerns. We immediately stood up a dedicated 800 number for COVID-19 questions and redeployed people from other teams to staff the call center. But it became apparent very quickly that we needed to do more to keep up with the call volume.

For perspective, normally, we would have 300,000-400,000 visits to our website each month. In comparison, we had over 1 million visits to our website within the first two weeks after COVID-19 cases started affecting the U.S., and a 40% increase in calls to our contact center. That’s why we started to look at our digital communication channels for a solution to potentially integrate into our systems for a more efficient triage process. That led us to deploy Hope, our AI-powered chatbot, built on Service Cloud.

A chatbot brings Hope to healthcare

We deployed our chatbot — with its straightforward point-and-click interface — within 24 hours. It’s worth noting that it was incredibly important to us that we didn’t sacrifice compassionate consumer engagement for expedient service. That’s why the team customized our chatbot with our brand colors and scripted the answers to frequently asked questions the way an empathetic practitioner — not a robot — would answer them. We also named it “Hope” to help it exemplify our mission.

Now, when people come to our website looking for COVID-19 information, they are given the option to immediately start their search for information with the chatbot instead of navigating through the site or calling the service center.

We gave Hope, the chatbot, a simple menu-driven script to help our customers with three main calls to action:

• The CDC COVID assessment — any “YES” answer drives the user to download our app for a telehealth screening
• Disease information — provided through the website and in-chat messages
• Additional support — how to contact our COVID-19 hotline

About 35% of our website visitors engage with the chatbot in some way. We estimate that each of our 250,000 chat sessions per month has been able to deflect one call to the service center while triaging patients appropriately. This, in turn, has allowed our staff to focus on the personalized care and resources for the patients who most urgently need it.

Three key learnings from this process

Going through a rapid response exercise like this has been an incredible learning process to help our internal teams understand how our patients want to consume information and how they want to communicate with us.

The key to this has been the integration we have between our texting tools, patient records, the chatbot, the CRM, and the other tools we’ve used through the App Exchange. This has really allowed us to be flexible and meet consumers where they are and how they wanted to be communicated with, and when.

The three big takeaways from this have been:

• Take time to personalize. By naming our chatbot Hope, we made it feel more personal and reflect our mission of Extending the Healing Ministry of Christ. We also personalized the script, so it felt like talking with a caring nurse, not a robot. This made patients feel more comfortable interacting about such a sensitive topic.
• Map out use cases. Each script is based on a FAQ or patient need with an associated case management workflow associated with it to either help the patient learn more on their own or be directed to the right team to help. This is critical to deflect calls.
• Integrate your communication channels. The chatbot is aware of the other communication channels. Instead of the patient calling into our call center and saying, “I’m looking for my COVID test result,” we were able to text them immediately when their test result was posted. If they asked the chatbot about it, they’d be directed to log into the patient portal and get their results immediately to avoid a call to the contact center.

High touch and high-tech coexisting productively, for your health

We’re planning to iterate on Hope after the pandemic subsides to make some form of self-service triage available to our patients moving forward. We’re also feeding more data into Health Cloud so we can automate more information proactively to our patients.

As a healthcare system, we have to meet our current patients and their families where they are. They’re looking for apps. They’re looking for websites. They’re looking for self-help and self-guided ways to engage with our healthcare system.

Continuing to improve in that regard will help solidify the brand loyalty that makes them want to come back to us again and again for their healthcare. All of which goes back to our mission and promise to help our patients feel whole through compassionate care and world-class expertise.

Learn more about chatbots on Service Cloud.