Artificial Intelligence (AI) is redefining how customers interact with insurance providers. There are more than 50 AI insurance use cases already in operation and nearly one in five insurers has already implemented at least one AI technology, according to McKinsey. By enabling AI to handle simple (and even more complex) customer interactions, insurers have improved the speed and accuracy of their services. Additionally, by arming sales and service staffers with AI as a digital colleague, insurers have evolved their workforces into always-ready concierges who can assist customers no matter the subject or issue.
At IPsoft, we’ve worked with many brands to implement our cognitive AI solution, Amelia, according to their specific needs. What we’ve learned is that not every AI implementation should be handled the same way. Digital colleagues, like humans, should act in a manner that’s appropriate to their respective task. Amelia as a personal shopper for a clothing retailer can make small talk, joke around, and have a more light-hearted disposition than Amelia would as a kiosk-style assistant checking people into a healthcare clinic.
Given the wide variety of use cases in the insurance industry, organizations must be cognizant of how best to deploy AI for customer- and employee-facing interactions. Whether Amelia serves as a whisper agent, support rep, an insurance agent, or even as a health insurance agent, she’s designed and programmed to look and perform to exact specifications. This allows insurers to be confident their companies are in compliance with laws and regulations, and that customers are never turned off by a bad experience. In this post, we’ll examine three of IPsoft’s insurance engagements and how they can serve as best practices for future implementations.
The Whisper Agent
Allstate, the largest publicly held personal lines insurer in the US, first deployed Amelia in September 2017. She has collaborated with Allstate’s live agents on more than three million calls, leading them through step-by-step procedures on a variety of support issues, including policy details and policyholder information. Amelia is trained on almost 40 different insurance topics. She has lowered call duration from 4.6 to 4.2 minutes, and 75% of customer inquiries have been solved during the first call compared with 67% prior to Amelia’s hiring. In one month alone Amelia assisted on almost 250,000 calls.
The key to this type of implementation is that Amelia guides human employees toward resolutions. In this sort of scenario, she’s a behind-the-scenes associate who can dig up information that may otherwise have been buried in dense manuals or on hard-to-navigate web pages. The difference here is that Amelia can find this information in a hurry. Rather than scrambling for information while customers wait on hold, human agents can type a question to Amelia and she can provide that information immediately, which is then passed onto customers.
Policy Questions and Escalation to Humans
MetLife hired Amelia to field questions from existing and prospective customers relating to policy changes, quotations, payments and pricing. Amelia is accessible from MetLife’s homepage and is currently integrated with a third-party web chat tool in order to promote human escalations for unique or more complicated matters.
Digital colleagues should always be viewed as collaborative solutions, rather than as human replacements. In a scenario similar to what MetLife has done, companies will be able to offload some of the more mundane and repeatable tasks to Amelia, such as policy cancellations, changes and payments. However, when a new request comes in that can’t be automated, it’s critical for a digital colleague to work with a human agent to ensure that the customer has received the service they require. In many instances, when Amelia escalates a conversation to a human agent, she will listen in the background to see if she can learn and incorporate new skills into her repertoire. If she feels she can successfully repeat what the human agent has done, she will request that the new skill be added to her current roster of abilities.
Liberty Mutual Insurance set out to combine the personalized consultative experience of a human agent with the speed and efficiency of a web or digital application. The insurer had an ambitious goal to train Amelia on a complicated customer-facing use case that involves leading customers through 25 to 30 questions in as efficient and comfortable way as possible.
Consumer/agent interactions can be difficult for many customers, who often call insurance companies because of traumatic or painful circumstances such as a death or accident. Finding a digital colleague who can make this process as painless and seamless as possible is critical to success. You don’t want an AI system handling a life insurance claim to speak in the same tone as an AI system that is helping a customer redeem loyalty points. Amelia is capable of recognizing and empathizing with the most common emotional responses, including happiness, sadness, anger, excitement and fear.
These are just a few of the many best practices to consider when deploying AI for insurance use cases. In many cases, you’ll have to make decisions specific to your business, your audience and the products you sell to determine the best ways to facilitate human-to-digital colleague interactions. No matter what you choose to accomplish with your implementation, always foster collaboration and emotional intelligence, and you’ll be on the right track.