Chatbots are revolutionizing the insurance industry

The use of chatbots is growing exponentially across the economic landscape, particularly in industries like insurance where the customer experience is tied directly to the bottom line. In fact, since 2019, the use of chatbots as a brand communications channel has nearly doubled, so two-thirds of global consumers have encountered at least one in the past year.

But not all insurance companies are seeing the same return from chatbot technology. Unlike previous advancements in customer relations, chatbots are not a plug-and-play or set-and-forget type of platform. Rather, they must be targeted at specific needs within the customer-facing application suite, then carefully honed over time to account for changing needs and expectations from an increasingly diverse consumer base. If done wrong, chatbots can lead to frustration and downright hostility toward the firm, which not only leads to customer erosion but tarnishes brands on social media and in word-of-mouth circles in ways that can seriously inhibit future performance.

So while the urge to deploy chatbots quickly may be strong, insurance companies would be wise to take a moment to truly understand this technology and how it should be deployed in order to create a rewarding experience that produces the greatest value for customer and provider alike.

Robot helper

A chatbot is basically an intelligent digital agent that can help customers resolve basic, and perhaps some not-so-basic, issues regarding products and services. This is an advance over traditional voice-driven menu navigation in that the chatbot is able to query large amounts of data from many disparate resources to better assist the customer through the entire interaction with the company. Rather than limit queries to a few predefined menu choices, a chatbot is able to understand and respond to a far more diverse set of requests, and it can do so in a much more natural, conversational manner than a simple voice system.

It doesn’t take a lot of imagination to see how this can improve customer relations, particularly in an industry as complex as insurance. For one thing, the often lengthy wait for a live customer service representative becomes a thing of the past. In most cases, customer inquiries are of a routine nature, having to deal with payment schedules, changes to coverage and the like. Chatbots can not only resolve these issues faster than a live agent, but they can also quickly determine if the requests exceed their capabilities and must be passed to a human operator, which helps streamline even these more difficult, time-consuming challenges.

At the same time, this lowers costs for the company. With more queries being handled in an automated fashion, organizations derive greater value from both the chatbot platform and call-center staff because both will be handling the levels of difficulty they are most suited to address. And since the chatbot is able to parse through vast sums of data even as it is providing direct assistance to the customer, it can identify and correct issues that the customer did not even know existed and guide them to additional services they might be unaware of, thus generating both new revenue streams and intra-cross-channel marketing opportunities.

It also goes without saying that chatbots are available 24/7/365 and don’t take vacations, sick days or personal time. Nor do they draw salaries or require matching benefit programs or resources dedicated to resolving interpersonal disputes, issues regarding advancement and promotions or office space.

Chatbots in action

While chatbots may be a relatively new technology in the digital enterprise, they are already delivering substantial results in real-world production environments. Verizon, for one, uses them extensively to handle initial customer support inquiries and is striving to leverage both human and technological resources to create a “human connection” with its user base. To date, more than half of respondents are comfortable with fully automated interactions, while more than three-quarters rate blended human-technology services as superior to those that are human-only.

The Transportation Security Administration (TSA) is also using chatbots and social media to reduce check-in times without compromising safety and security. The agency leverages them primarily to respond to routine questions on topics like prohibited items and screening procedures. In this way, answers can be delivered within two minutes as opposed to the one-and-a-half hours of a human-centric system, which is usually long after the flight has departed with the passenger onboard.

Meanwhile, Bank of America has made its service, named Erica, available to more than 19.5 million customers and has logged upwards of 100 million interactions with users, reporting more than 90% efficacy for useful answers. The company reports that the platform’s success is based on two key factors: The ability to answer basic questions quickly and accurately, and the ability to engage in direct conversation just like a teller or call center agent would.

Of course, to achieve this level of proficiency, chatbots require some fairly sophisticated technology under the hood. First, it needs support from a large-scale, robust infrastructure in order to access the data it needs to deliver effective service; but this is more than just server, storage and networking technology. Substantial resources must be devoted to data conditioning and preparation to ensure the system can interpret information in the correct context and is not influenced by bias, inaccuracy, irrelevancy or a host of other limiting factors.

Talking the talk

Another crucial piece of the chatbot puzzle is advanced language skills. This relies on a particular form of artificial intelligence called Natural Language Processing (NLP), which uses sophisticated algorithms and high-speed processing to provide the kind of conversational interaction that users say is the most engaging. Recently, NLP has evolved with new forms of symbolic understanding (symbolic NLU) that more closely match the way the human brain develops language skills and interprets the intent of language, not just the literal meaning of words.

Other forms of AI, such as machine learning (ML), as well as more basic technologies like keyword recognition and standard logic trees, are being used to create a wide variety of chatbots, each suited to a particular role in both customer-facing and back-office applications. Some bots, for instance, only need to provide limited assistance based on pre-set query options and data sets, while others can learn from their interactions to provide increasingly sophisticated responses as their data environments evolve.

Perhaps more than any other technologies entering the workplace today, chatbots offer the most promise when it comes to digitizing enterprise processes for the 21^st-century economy. Today’s business environment moves at the speed of electrons, which means customer interactions need to be lightning-fast and universally rewarding – and this all must happen at the extreme scale of a global marketplace.

Trying to accommodate these needs with a strictly human workforce is both prohibitively expensive and largely ineffective considering the sheer number of queries coming in and the vast amounts of data that must be parsed to produce a satisfying result. Machines like chatbots will never replace the human workforce, but they do wonders to augment their skillsets to engage with customers and provide timely and effective resolutions to their problems.

And as has always been the case in business, ensure a happy customer remains a loyal customer.

Pamela Negosanti (pnegosanti@expert.ai) is global head of insurance at expert.ai.

Related:

• Insurance inclusion means remembering the non-digital natives
• How to solve the insurtech distribution problem
• Human touch is key in claims automation