Lincoln Financial's chatbot part of larger omnichannel experience investment

Lincoln Financial Group (Radnor, Penn.) has launched a chatbot, "Ask Abe," to help customers handle routine questions on its website. The life insurance company first launched click-to-chat on its site, to connect with live representatives, in March 2017; these service reps remain available during normal business hours for help with more complex tasks.

Jay Mahapatra, AVP, omnichannel delivery says the carrier is working to shift more capabilities onto Abe over the next six to 12 months, for both policyholders and agents .

“Policyholders will be able to make a payment or change a beneficiary through micro apps,” Mahapatra explains. “A new window will open and we’ll take the user through a simple step-by-step process to perform tasks on their own.”

Bob Scheppegrell, SVP of customer solutions, Lincoln Financial Group, says that Abe is pare of a broader digital strategy at the company that encompasses three key pillars: customer experience, increasing operational efficiency and innovating with digital and data.

"One of the key components is our omnichannel communication, and our chat and chatbot fit within that part of the strategy,” Scheppegrell says.

di-lincoln-lfg-061818.jpg

Lincoln Financial built Abe internally, through multi-departmental partnerships. leveraging the Google DialogFlow platform to create a robust conversational experience, the executives say. Development took three months, from inception in September 2017 to the December 2017 pilot launch. Customers type in a question, which is sent to DialogFlow. The platform then provides the customer with corresponding responses to the generic questions on the chat window in real-time. Lincoln Financial built the solution through close partnerships with internal stakeholders.

Right now, Lincoln Financial is continuing to build out Abe’s capacity for answering questions and expanding its vocabulary of recognized words and phrases, an ongoing process. Nine months ago, the insurer began to train the bot with 17 simple questions asked in a number of different ways, like how to make a payment or change a beneficiary. Now, Abe is capable of addressing 33 scenarios and can provide more than 500 varying responses to the basic questions.

"We created these capabilities by mining data collected during the pilot,” Mahapatra explains. “We collected the chat history from the time we launched the click-to-chat capabilities in March 2017 to September 2017. We identified the top questions our customers ask our customer service reps and mined that data to come up with the best questions, scenarios and different variations.”

The company will continuously train Abe with more questions based on the questions customers continue to ask, using analytics to analyze customer feedback captured from the simple chat and chatbot. “Eventually, the system will self-learn, but right now, we take data on the questions customers ask each day that Abe can’t yet answer and train the chatbot to be able to respond appropriately,” Mahapatra says. Currently, Abe can answer 85 percent of customer queries, but Lincoln Financial is working to improve that rate to 99 percent by the end of the year.

One challenging aspect of the chatbot is its inability to provide empathy. Bringing in a human aspect to the chatbot is critical, especially when dealing with life insurance claims and beneficiaries’ losses, so Scheppegrell says a key objective of all human and electronic communications is responding with empathy. “We want to make sure we extend that to Abe, so as Abe communicates with customers, they get the same feel that they’re being taken care of that they would if they were chatting with a human,” he points out.

As Lincoln Financial builds out Abe’s capabilities and the chatbot begins to teach itself and answer more questions, the company will pilot it within its life insurance division and then intends to expand to other businesses, including the annuities line of business.

For reprint and licensing requests for this article, click here.
Artificial intelligence Customer experience Customer service Machine learning
MORE FROM DIGITAL INSURANCE