From the Data and AI Expert Labs: Break open your processes
Today, I want to reinforce that you shouldn’t rely on a compelling incident. You don’t need to wait until a system breaks to make it better. A great example is the Interactive Voice Response systems (IVR) that are pretty standard in customer service. We’ve all interacted with them. They’re slow, they’re dumb, and we’ve been stuck with them for years. IVR is one of the few technologies its users actively seek to circumvent. But up until the public health crisis of COVID-19, IVR just wasn’t broken enough to justify a replacement. But during the crisis, IVR was too broken to cover for the shutdown of staffed call centers. AI and voice recognition technology became practically a requirement for success.
Technology like conversational AI, machine learning and natural language understanding give organizations more options and flexibility. And you don’t even need to ditch your IVR system if it’s still working for you. For example, Cardinal Health, an American healthcare company, discovered their IT agents were fielding many password reset requests. Users needed to reset their password without access to their computer, so they had to call in the issue. Usually, they would need to navigate an IVR menu and speak with a human agent, who then must then go through the manual process of resetting the user’s password.
Cardinal Health understood two things about this situation: users do not like deviations to their routine, and they want a solution fast. So Cardinal Health used a Watson Assistant for Voice Interaction service to embed conversational AI into their IVR system. Now, when users need to reset their password, they connect to the AI, and the issue is resolved immediately.
Businesses can now create different and unique experiences for multiple channels while still maintaining a consistent brand voice. If a customer is having a bad experience with IVR, they can switch almost instantly to a channel they prefer. Customers often fit into two base types. One doesn’t ever want to talk to a machine; the other wants to handle their problem without human intervention. An AI solution allows you to serve both customer bases without detriment to the other.
Now, organizations have a tremendous amount of data, and that data will form the building blocks of what your solution can do. You may have a system that, on the surface, seems like it works for customers, but that doesn’t mean it can’t be improved. Take the example of Geico, who proactively improved their insurance application process. Customers had no problem filling out an application. But the second-largest auto insurer in the United States introduced a virtual assistant to ask the customer a set of basic insurance-related questions. As the customer answers, in the background the AI completes their insurance application. All in a very conversational manner. Once the conversation is finished, all the customer needs to do is submit the application.
All of these use cases got started using what they already had. You can do the same. During the current crisis, customer requests for information have grown so large that no team of humans can handle it all. To me, that’s a broken system. And it’s simply unfeasible to replace all of your human workers with AI. No matter how advanced the technology is, it’s not a suitable replacement for the human. Instead, you need to gather your team and figure out how to make the system better.
In my final post of this series, I’ll talk more about how this technology will affect your team and how you’ll need to invest in the human, more than in the technology.