As customers increasingly demand a better experience when they interact with companies, including insurers, help is coming from a counterintuitive source. It turns out that one of the best ways to be more personal is through… robots.
More precisely, the answer is turning out to be chat robots, or “chatbots.”
People don’t like having to phone call centers and wade through that phone tree — “Para continuar en espanol, oprima uno… For billing, press 2; for….” Many, especially younger people, just want to be able to text a question and get it answered. That’s how they handle everything else. So, many companies are realizing they need to have customer service reps that respond to texts, and they’re seeing an opening to use chatbots.
Using so-called natural language processing to understand a text message and then drawing on artificial intelligence to both find the answer and generate a reply, chatbots can handle perhaps 70% to 80% of queries. They can hand a conversation off to a human when necessary and can take the conversation over again, without the customer’s ever realizing that a bot has been involved or that a handoff occurred. In fact, the bots can wind up sounding a lot less robotic than the standard call center rep who is only allowed to read off a script.
The bots are so efficient about finding answers that they actually have to be slowed down, so the customer doesn’t think, “No one could type that fast,” and wonder if a computer is involved. (A certain percentage of typos can also be programmed to appear, as can emojis or lots of exclamation points, to make the bots seem more human. You can actually program the bots to have different personalities.)
With so many mundane tasks handled by bots, the call-center reps get to deal with more interesting issues and can spend more time with customers, giving everyone a better experience.
Although they haven’t shown up widely in insurance yet, they are in use in numerous other industries, with great success.
Chatbots have been around for more than 20 years. Why should companies pay attention to chatbots now?
For starters, these days just about everyone is carrying a super-computer around in her pocket. In 1991, 1GB of flash memory would have cost around $45,000. Now, most phones have at least 32GB of memory. Processors are more than 1,000 times as fast as they were in 1991. So, the technology for chatbots is lightyears ahead of where it was.
Companies have also placed an increased focus on messaging, including with bots. Facebook Messenger uses more than 11,000 chatbots to respond to messages. The chat app Kik recently said that more than 20,000 bots have been made specifically for its platform.
Perhaps even more importantly, the pendulum in the customer-business relationship has swung heavily in favor of the customer. Companies no longer control the message/brand; it’s all out there in the ether, and companies need to guard their reputations by caring for customers. Some companies, such as Zappos, have pretty much built their businesses on the customer experience, while others, including cable companies (Comcast, most notoriously), are vilified.
Insurance companies can see what’s happening in other industries and see what they need to do. Net Promoter Scores (NPS) are the insurance industry’s most consistent measurement of customer loyalty, and, despite some pockets of brilliance by individuals, most of the time the insurance industry stinks. Chatbots can help.
Chatbots create a conversational web and conversational commerce. They can even be programmed to wish the customer a happy birthday or happy anniversary or make some comment about how long the customer had been with the company.
Chatbots make a company’s behavior more consistent across the board, especially in terms of elegance and simplicity. On top of that, it’s easy to keep the bots on-message, and they only need to be trained once.
See also: ‘Age of the Customer’ Demands Change
In 2013, it was estimated that it takes five screens to get a user to where he wants to go. In 2015, that number has increased to seven. Bots get the information almost instantly, even if that means going to a deep link in an app or on a site or in a corporate data center.
Bots aren’t “one size fits all.” They’re “one size fits one.” So work has to go into customizing them for a company. But simple bots such as for frequently asked questions can go live in a day, and an ecosystem of bots can be developed over time. Once a bot exceeds its ability to answer a question, it might initially pass the question to a human, but, in time, the handoff could go to another bot that has been developed.
Bots will also be able to take on more tasks, including outreach to customers. Bots could automatically alert customers of an impending hurricane and begin a dialogue with them about what steps to take to prepare, who to call if they need immediate assistance, how to file claims, etc.
In insurance, there’s nothing like the Domino’s pizza tracker, which allows a customer to follow along with an order every step of the way, from the order to the oven to the front door. But there will be, and imagine how helpful that will be with claims. Many customer calls are about where their claims are in the process, so streamlining it and providing a bot with the capabilities to respond to the customer would make the process easier, eliminate a lot of calls — and make the customer much happier.
Of course, an insurance bot isn’t going to answer “what is the meaning of life?” or “how much wood could a woodchuck chuck if a woodchuck could chuck wood?” like Apple’s Siri can. But a bot can be tailored and trained to answer many questions, filling a gaping hole in the insurance world.