In the insurance industry, some of the most important calls come under some of the most stressful circumstances. That’s why CodeObjects is using IBM Watson® technology to develop InsurBot.ai, an AI assistant that eliminates call center holds, reduces costs and helps put anxious customers on the road to recovery quickly.
Insurance calls surge after natural disasters, creating backlogs and long waits for policyholders. CodeObjects wanted to use AI technology to help insurance companies respond to calls more quickly.
InsurBot.ai helps insurance companies improve responsiveness by automatically responding to policyholders’ questions and requests, and automating business transactions.
The years 2016 and 2017 saw record numbers of natural disasters in the US and beyond. For property and casualty (P&C) insurers, this meant record numbers of catastrophic loss claims from hurricanes, floods, fires and more. As the number of claims grew, so did the need among P&C insurers to provide impeccable customer service. This was a call to action for CodeObjects, which develops core insurance systems for insurance companies throughout the US.
“If someone suffers a loss from a hurricane that damaged their house, they don’t want to spend an hour filing a claim with their insurance company,” explains Arun Bala, Vice President of Product Management for CodeObjects. “The only interaction many policyholders have with their insurance company is when they need to file a claim, so it’s in the insurer’s best interest to make the customer experience as positive as possible.”
However, many insurance companies outsource this vital touchpoint to third-party call centers, which can create complications. Outsourced call centers typically use their own systems because they don’t want to train their staff on several different applications. This can impact an insurance company’s ability to obtain accurate information about the nature of the loss, which in turn creates inefficiencies in the entire claim handling process.
“And it’s not just calls related to catastrophic events,” says Bala. “People call to find out what sort of coverage they have, or to ask questions about their bills. Every single touchpoint they have with customer service reps can be fraught with inefficiency, not to mention the high cost. These are significant pain points, and many insurance carriers are trying to mitigate them using automation.”
“In response,” says Bala, “we are augmenting our core system offerings with chatbot capabilities that will allow our existing and prospective customers to provide a superior customer experience.”
“We believe chatbots make a compelling case. Once you train the model well enough, it provides predictable and repeatable outcomes. Once you train the chatbot to converse like your five-star CSR [customer service representative], every single conversation has the potential of being a five-star conversation,” says Bala.
After exploring a number of options for AI, CodeObjects selected the IBM Watson platform to support its chatbot product, which the company calls InsurBot.ai. The solution features IBM watsonx Assistant, IBM Watson Discovery, IBM Watson Speech to Text, IBM Watson Text to Speech and IBM® Voice Agent with Watson services.
Anthony Peccerillo, Vice President of Product Marketing for CodeObjects, picks up the story: “Our plan for phase one of the product was to target the customer service area and automate repeatable tasks that people are doing on a daily basis.”
Critical first-phase tasks included filing a first notice of loss (FNOL), checking payment or claims status, getting a quote and making a payment. The solution responds to policyholders’ voices, following a predetermined script that was developed after using Watson technology to analyze approximately 50,000 calls received after Hurricane Irma in 2017.
Although the intention was to make the experience as easy and seamless as possible for policyholders, InsurBot.ai was also careful to give callers a sense of control. Bala explains: “In the dialog flow we wrote using IBM watsonx Assistant, there are various nodes where the call could be branched off to a live agent, either because of user request or because of some other internal condition.”
Just as CodeObjects was finishing up the scripts for the first phase of the InsurBot.ai initiative, Hurricane Michael made its first appearance on radar screens. The hurricane was headed for the Florida Panhandle, where many of CodeObjects’ customers had operations. If the models were correct, it would be one of the largest storms to hit the area since 1992. Catastrophic losses were virtually inevitable.
Anticipating a huge influx of P&C claims after the hurricane hit, Security First Insurance, one of CodeObjects’ existing customers, decided to roll out the new AI assistant in advance of the storm. With less than a week’s notice, the InsurBot.ai and Security First Insurance teams got to work on moving from pilot to production. “We went into deployment mode,” says Bala. “It was all hands on deck. We tested it, stabilized it and migrated the production environment from standard to premium Watson plan. It was a short window for a lot of activity, but we met the challenge head on and had a successful outcome.”
Werner Kruck, Chief Operating Officer at Security First Insurance, agrees: "Not only are we able to improve customer satisfaction, the team at InsurBot.ai enabled us to create a technology strategy to further differentiate the experience we offer and lower costs in the long term.”
Among the policyholders who chose to use InsurBot.ai after Hurricane Michael, 75 percent were able to complete their business without being diverted to a live agent. In total, InsurBot.ai handled 30 percent of all of the calls Security First Insurance received after Hurricane Michael’s devastating landfall. This resulted in savings of approximately USD 1 per minute during the post-storm spike.
Furthermore, in keeping with CodeObjects’ tenet of helping its customers create frictionless customer experiences, callers who chose to use InsurBot.ai experienced zero wait times. During a hurricane the previous year, the average wait time was approximately eight minutes, and some calls took as long as 30 minutes. Clearly, this was a huge improvement for the insurer—and for its policyholders.
The success of InsurBot.ai during Hurricane Michael confirmed for CodeObjects that its solution will help address customer service concerns throughout the P&C industry. Ultimately, InsurBot.ai will support multiple languages and multiple channels, plus it will integrate with many existing insurance systems.
CodeObjects plans to use analytics to round out the solution further. Peccerillo explains: “For example, we could overlay data that’s projecting the path of a storm with the client list and proactively send risk mitigations, documentation or videos so people who are in the track of the storm can prepare.”
Analytics could also benefit the insurance companies themselves. Says Bala: “We want to provide our customers with tools that tell them what’s going on. What’s working well with the bot, what’s not working well, and where the hotspots are so that they can quickly address them.”
With these additional capabilities, CodeObjects believes that InsurBot.ai will be able to handle nearly all calls seamlessly, even during catastrophic events. “It’s designed to handle the highest volume you could possibly throw at it. Whether you’re the first person or the 100th person, your call will be answered immediately. That’s our goal,” Bala concludes.
Founded in 2006, CodeObjects develops core systems for insurance companies across the US. In 2018, CodeObjects introduced its flagship product, InsurBot.ai, an AI-powered customer assistance service that is changing the way insurance companies communicate with their customers.
Legal
© Copyright IBM Corporation 2019. IBM Corporation, IBM Watson, New Orchard Road, Armonk, NY 10504
Produced in the United States of America, January 2019.
IBM, the IBM logo, ibm.com and IBM Watson are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the web at “Copyright and trademark information” at ibm.com/trademark.
This document is current as of the initial date of publication and may be changed by IBM at any time. Not all offerings are available in every country in which IBM operates.
The performance data and client examples cited are presented for illustrative purposes only. Actual performance results may vary depending on specific configurations and operating conditions. THE INFORMATION IN THIS DOCUMENT IS PROVIDED “AS IS” WITHOUT ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING WITHOUT ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND ANY WARRANTY OR CONDITION OF NON-INFRINGEMENT. IBM products are warranted according to the terms and conditions of the agreements under which they are provided.