The current boom in artificial intelligence technologies is currently at the forefront across many industries – insurance included. With so many use cases already defined and proven, such as underwriting and better customer interactions – as thoroughly explained by Verint North Asia and Korea vice president Matty Kaffeman in the first part of this Insurance Business Asia interview – there are still areas left to be improved.
“If you look on what we call the engagement capacity gap, we have a boom in interaction – customers can talk with the organization in many ways, either through voice or digital. A customer’s expectation is tied to the best service. If it’s not there, I'll just move to another insurance company,” Kaffeman said in conversation with Insurance Business Asia.
However, while this technological surge has proven to be useful, there are still gaps that companies cannot easily plug, all dependent on their individual capacities. Kaffeman puts this down to not enough people, the associated workload, less capital, or any combinations of the three. This gap, he said, can be helped by AI, and it all comes down to how smartly a company can apply it.
if you look at the major cost that businesses invest in the contact centre, it’s close to $2 trillion the cost of the workforce,” Kaffeman said. “At Verint, the solutions that we offer do not cost as much. In fact, any solution you invest [in] will help you. But here, what we offer is a platform. So, we have a platform that takes the data from all the different silos. That could be a chat system, an email system and voice system, a webform – anything that you have that can be placed into one data hub that allows us to normalize the data.”
One of the more usual problems that any growing company can run into is the localization of data. Kaffeman noted that with different systems in place, it can sometimes be inefficient to figure out where the information you need is stored or who is the right person to ask, whether you are an employee yourself or a customer.
“Another thing that we use AI is for example for knowledge management. There’s a lot of questions and answers circulated in the system that we capture, whether it's by voice whether it's by a text. We have a bot that will analyse them, will go through them and find out what are the key topics or what are the key knowledge articles that we need to update so that customers can be better informed,” Kaffeman said.
“For example, during COVID, the guidelines provided are not sufficient, so what you get are people asking too many questions about it and sporadic content that does not always provide the right answer,” he said. “With AI applied a human agent that is responsible, rather than trying to craft answers on the spot, can just look for an option, suggestions, draft around that and use it. In a sense, I'm not replacing the people, but I'm empowering them.”
Encountering these issues more often than most is the management suite, Kaffeman said, and this is a particular area where the power of AI can truly shine. Rather than risk inefficiency in sifting through different systems to get the right answer, having automated solutions that can come up with something right for every single use case makes more sense.
“Another role that benefits from the AI is the management suite,” Kaffeman said. “If management without AI tools try to understand or try to pull out information, a lot of times it will come from different sources, different silos that we have in the organization. They also have to define the data that they want to use, they have to define the question they want to ask. With AI, you can have bots that are trained on interaction data that allows them to identify abnormalities.”
Powerful data delivery aside, Kaffeman also noted its potential to provide the information in a way that can be more easily understood. Highlighted anomalies can easily be translated into neat visualizations – as he mentioned, rather than words, you can have AI replace it with graphs for managers to easily connect to and respond accordingly.
“That’s something that if you try to create it today without any automation, you will have to figure out what to look for, you will have to find who can bring the data from different repositories,” he said. “This data has to be continuously updated into a place that can analyse the information consistently and understand qualities, both of which are key for organizations to be able to react.”
What are your thoughts on this story? Please feel free to share your comments below.
