New technologies generate a great deal of excitement, and there’s a positive buzz right now about the potential impact of artificial intelligence, machine learning and automation in the insurance industry. There’s little doubt that AI is set to play a major modernizing role, but carriers are right to be a little cautious about precisely how they embrace new systems.

Chatbots have many interesting applications in insurance, and we’re going to delve into some of them in a moment. However, there is no substitute for years of experience and professionalism with a sensitive touch. Certain situations will be better served by humans, and for AI to improve, human guidance is essential. Let’s take a closer look at when chatbots should and shouldn’t be used in insurance.

Handling Basic Queries

If someone wants to know if their child is covered on their policy, or what the excess is for a particular type of damage, a chatbot can provide the answer immediately, with minimal fuss, no matter the time of day. Basic, straightforward queries are ideal for chatbots. They can also address concerns and complaints by logging issues and gathering information, but there is a line with customer service where a human should step in.

With the benefit of natural language processing and sentiment analysis, a chatbot should be able to determine when a customer needs to be escalated to a human adviser. This is absolutely vital if you want to ensure that your customers don’t become frustrated. Chatbots are not equipped to deal with complex situations and delicate emotional states. They must be able to recognize when a customer is upset and quickly flag the need for a human adviser to step in.

The Right Level of Coverage

It can be difficult for customers to decipher all the different policies on offer and find the ones they need for real peace of mind. Walking them through what’s on offer and making valuable suggestions is time-consuming. Different sales agents will likely have different opinions. This is an area that’s perfect for chatbots. Consider that 74 percent of consumers told an Accenture study that they’d be happy to get computer-generated insurance advice.

Chatbots can quickly and accurately assess a customer’s needs based on all the information they have provided. They can be configured to ask pertinent questions, can cross-reference in real time and will provide consistent responses. Once fully tested and properly configured, chatbots can respond to leads, sell products, assess existing coverage and make valuable recommendations.

Supporting the Claims Process

Straight-Through Processing (STP) is driving greater automation of claims, which reduces errors, costs and the time it takes to settle. Chatbots play a central role in this process by enabling customers to file claims on their smartphones, gathering the precise data that’s required to resolve a claim and allowing customers to check on the status of their open claims.

For straightforward claims, AI also can handle fraud mitigation and even authorize and send payment. The purpose of the chatbot is to gather the necessary data and answer customer queries. An algorithm will assess the data automatically and decide whether a human agent is required. As we’re still in the relatively early days, human agents will be called in to oversee some aspect of most claims, but through machine learning the automated models will observe how claims are handled and use the information to improve. Over time, more and more claims will be fully automated.

Customers Prefer Chatbots

Messaging is the preferred channel for customer service in the U.S., according to a recent Twilio survey. The vast majority of people prefer messaging or email over face-to-face interactions with businesses, and phone calls are far less popular with younger generations, particularly millennials. Good chatbots quickly are becoming indistinguishable from people on messaging platforms, but it’s important to remember that there always will be situations when human advisers are required.

The key is in identifying when a chatbot can’t meet the needs of the customer before they become annoyed and have a bad experience. As a complementary technology, chatbots are an excellent way of freeing up your skilled staff to focus on issues and challenges where they can really add value and exercise their expertise.