Without giving anything away, Game of Thrones is about a bunch of humans fighting each other for power while a much bigger threat (the White Walkers) looms on the horizon. While not quite as dramatic, there are similar battles going on between humans and technology right now, often losing sight of the fact that we have common enemies to fight.
This is particularly true when it comes to customer experience. Technology, particularly artificial intelligence and chatbots, have threatened the long-established world order of solely using humans to help consumers interact with companies. What many companies have lost sight of, however, is that there is a much bigger enemy out there – bad customer experience.
Without belaboring how bad most customer experiences are now, suffice it to say that many companies aren’t living up to customer expectations. Consumers expect that companies will know who they are and what they want, and they will not forgive poor customer service that comes across as not valuing customer’s time.
Unfortunately, as companies began deploying chatbot solutions, many made the mistake of not understanding these truths and not investing in understanding their customer’s intent. They put the wrong metrics in place and didn’t focus enough on customer satisfaction. The experimentation phase is over and companies are now taking a much harder look at chatbot technologies. So, what does a chatbot comeback look like? The answer may lie in human agents.
New York Times columnist and best-selling author Thomas L. Friedman wrote a brilliant article on how humans and AI are working together to do this today. The article is based on Friedman’s recent visit to the 7.ai Bangalore office, where he interviewed agents about how our AI technology is improving their workflows. He saw firsthand how the world has changed—specifically how chatbots and humans are working side-by-side to deliver a personalized, predictive and effortless customer experience for our clients and their customers.
Investing in the right blend of AI and human intelligence (HI) can result in a “near human” experience for consumers. Chatbots cannot completely replace humans, at least not yet, and they have to work with humans to deliver the right results. AI investment is necessary because it is through the power of AI and machine learning that we can decipher huge amounts of data to truly understand what customers want. Only by understanding what they want can companies deliver the best experience to them.
A recent Forrester survey states that 46% of companies said sales and marketing are leading the investment in and adoption of AI systems, however, it is equally important to invest in the collaboration between AI and humans. The future of customer service is one in which chatbots powered by AI and HI are delivering convenience and customization that consumers look for—and provide the comfort of a humanized conversation.
There’s no part of the business where virtual agents and real people can work together better than in helping customers. Bots are fast and accurate. People are empathetic and have judgment. Together, they’ve got what it takes to deliver the best service possible. Based on my company’s research, we’ve identified three basic modes in which virtual and human agents work very well together.
When a customer gets on the phone with a virtual agent or chats with one, there’s always the possibility that the customer asks for something that the virtual agent can’t provide. That might be a complicated case the virtual agent can’t puzzle out. Maybe the customer indicates that they’re upset (which virtual agents can often spot easily, because the customer uses exclamation points, starts shouting or curses). It might even be a case where the virtual agent hands off to human agents by design.
Bots can easily collect information like the customer’s name or account number and a description of the problem, then suggest resolutions before handing the customer off to a person. (Touch-tone interactive voice response systems do this already, but bots are more efficient at it—and less annoying for the customer to converse with.) The human agent in this situation has a head-start; she already knows who she’s talking to and what the question is. And the virtual agent can weave in more information, like past history of product purchases, a service level, or a slot in the company’s hierarchy of loyalty levels. For example, if a hotel customer service agent knows the customer on the line has stayed at the chain twelve times already this year—and has been a loyalty member for 15 years—she can more quickly resolve a reservation problem and keep that customer happy. Having the customer’s background teed up by the virtual agent shortens the time to resolution and delivers a better experience.
Some companies aren’t comfortable with their customers talking to virtual agents because they want to provide a human touch from the start. One senior enterprise architect at Marsh & McLennan told Forrester Research in its December 2016 report on AI in financial services: “We would never have a wealth management client interact with a robot.”[i] But the agents themselves can become smarter because a bot is whispering in their ear.
Bot-assisted customer-service agents appear much smarter than agents who work alone. If there’s a new discount, the bot brings it to the attention of the agent. If something’s out of stock, or a new rule prevents the customer from accessing an offer, the agent will know that, too. It’s like having an invaluable assistant figuring out the best things for you to say to the customer.
If a new employee were working with customers on the phone, you might want to put a more experienced staffer on the phone at the same time, to look over his shoulder. That way, the more experienced worker could step in if the new guy gets stuck. That’s exactly how supervised bots work. A human agent might supervise eight or ten chatbot conversations—far more than a human agent could handle if they were doing the talking. When a virtual agent gets stuck, the more experienced human can step in and solve the problem.
But the human has one more job: to tag the customer’s intent, which the virtual agent may have missed. This information gets fed back into the virtual agent system and makes it a little smarter, so next time it can handle a similar interaction on its own. This is how supervised machine learning works, and over time, should reduce the level of human supervision. In this mode of interaction, the virtual agents do what they do best: handle routine cases quickly and efficiently. And the humans do what they do best—solve more complex problems, help people who need a little more empathy, and train the virtual agents on cases they’re not yet ready to handle.
As ironic as it sounds, the chatbot comeback will come as the result of better integration with human agents. The chatbot spring is coming, and with it, we will see a new crop of more intuitive, more human applications that better resonate with consumers.