BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
American Fidelity Assurance Company is using AI to eliminate menial tasks and improve the workflow for employees, as well as the customer experience.
At the MIT Sloan CIO Symposium, Shane Jason Mock, vice president of research and development at the health insurer, described how the company uses intelligent process automation -- a combination of robotic process automation (RPA) and machine learning -- to automatically read emails and prioritize customer needs to deliver faster responses to customers.
Editor's note: The following was edited for clarity and brevity.
How did your company leverage artificial intelligence in an impactful way in the past year?
Shane Jason Mock: We have built a robot to essentially read emails, and that's having real business value. We're using a combination of RPA, robotic process automation, and we're tying that to artificial intelligence, machine learning. The big win there is for our customers. They're getting their emails answered faster, things are getting routed through. It's helping our internal colleagues as well to be able to focus more on phone calls. It's been a big win across the organization.
How is security included in an intelligent process automation strategy?
Mock: Everything is very secure, we take security very seriously. There's a stringent security protocol there. Everything is on-prem, everything is behind a firewall, everything is encrypted. There are absolutely no security concerns from that perspective.
All customer emails were coming into a common queue. Some people are asking us, 'How do I file a claim?' Some customers are saying, 'I need to change my address.' Customers have all sorts of different needs. So we knew where everything got routed in the past. We essentially had a labeled dataset -- a supervised learning labeled dataset. We were able to take that dataset and train a machine learning algorithm to understand those emails and move those emails so customers could get answers faster.
Has American Fidelity explored the use of intelligent chatbots?
Mock: We've definitely looked at different chatbot solutions and we have some proofs of concept happening right now. Where all that goes, we're still in the determination stage.
As vice president of research and development, would you say hiring more data scientists or data engineers is happening at your company? If so, why?
Mock: The whole world is going toward data. I've heard lots of people say data is the new oil, so that's an interesting concept. We're having some tremendous success working with folks that have petroleum engineering backgrounds, electrical engineering backgrounds, actuarial backgrounds, so not the traditional data science background. But they have the ability and capability to be trained on the job and learn what they need to know. I think if you have the building blocks and an understanding of the deep math concepts, you can pick up the data science concepts pretty easily.