AI in service: What happens when 1+1=3?
It was an epic date to witness—Wednesday, October 21st, 2015—Marty McFly and Emmett “Doc” Brown arriving in a time-traveling DeLorean from 1985.
At least, that’s what happens in Back to the Future: Part II, which debuted in theaters 26 years ago. In the movie, Michael J. Fox’s iconic character, Marty, steps out of the time machine and into a futuristic 2015 filled with flying cars, hoverboards, self-drying jackets, and power-lacing Nikes.
From the perspective of the real 2015, most of the things depicted in the movie’s fanciful future still seem delightfully implausible. But one thing the film did get right is the use of artificial intelligence (A.I.) in 2015 as a customer service tool.
Customer service, digitised
In the film, when Marty enters a diner called “Cafe 80s” he is greeted by two computerized waiters appearing on a TV monitor in the form of Ayatollah Khomeini and Ronald Reagan. They’re clearly artificial yet not very intelligent, but they understand Marty’s increasingly frustrated vocal commands well enough to stop talking and serve him a Pepsi.
It’s a 2015 customer experience facilitated by A.I. technology, approximating the actual level of customer service one expects these days, here in the real world, from the likes of Apple’s Siri, Microsoft’s Cortana, and Google Now. These tools are increasingly clever and can answer basic queries, yet it’s hard to forget you’re dealing with software.
But then there’s the advanced “cognitive computing” software known as IBM Watson. With its natural-learning algorithms, Watson has been widely touted for its potential to deliver a groundbreaking customer experience that puts smartphone “personal assistants” to shame. In fact, some fear that it may can replace human customer service agents entirely—though maybe not just yet.
There are still some big differences between human customer service agents and even the most impressive forms of software. And these differences suggest that, for the time being at least, the most effective customer service solution will be one that manages to combine the best of both worlds. Here’s why:
1. Software lacks E.Q.
Watson is best known as the supercomputer that beat reigning Jeopardy! champs Ken Jennings and Brad Rutter in 2011, but its power has been put to more productive and practical use in the world of medicine, where it has helped researchers sequence cancer patients’ genomes and helped doctors develop custom treatment plans. Yet despite the considerable value it is delivering, nowhere is there a hint of doctors using Watson to interact with patients on their behalf, extolling the virtues of its superior bedside manner. And there’s a good reason for that: Watson may have an I.Q. to rival Stephen Hawking’s, but it still lacks emotional intelligence, or E.Q.
In many customer service situations, the ability to empathize with a customer and genuinely appreciate their frustration or desire is the key to providing world-class service. Watson may be the best thing since Google when it comes to efficiently resolving customers’ questions, but it’s likely to be at a loss when it finds itself in the deep end of a rapidly escalating or sensitive situation that demands calm and quick-thinking human intervention.
2. Humans lack processing power
Providing a great customer experience typically requires collaboration across an enterprise, with employees reaching across organizational silos to find the data they need to resolve an issue or guide a customer on the next step of their journey. This can prove challenging, however, in a world where gigabytes of data are being generated daily—from recorded phone calls to live chat transcripts to emails and beyond. Human beings are great at generating data, but sifting through it is another matter. Depending on the situation, resolving a customer’s question can sometimes involve numerous agents, working across separate departments, to piece together an answer—if they can mine the right data to begin with.
Of course, that’s one area where computers have a distinct advantage. IBM Watson is especially adept at making sense of vast amounts of unstructured data—data that would otherwise be difficult to query at all—to find key information, fast. Some of its best implementations in a customer service capacity have bridged the gap between FAQs and live agent chats, resolving customers’ queries almost instantly by drawing upon vast troves of information to deliver quick and meaningful answers. And providing the right response to a customer at the right time, with the right context, is a big component of a great customer experience.
3. Software and human agents together can make 1+1=3
Software operates on conventional either/or binary logic—input, output; if this, then that. In the mathematical thinking of even the brightest A.I., 1+1=2, and it takes a great deal of additional programming to convince a piece of software otherwise. But in the world of customer experience, sometimes a service agent must bend the rules to satisfy a customer—effectively making 1+1=3. And this is where the strengths and weaknesses of human agents and software can complement each other well, ensuring that 1+1=3 for customers more often than not.
IBM Watson excels at speedily processing information and producing helpful solutions, doing things that human agents simply cannot. But in any customer journey there are moments of truth when it’s important to get proactive, empathetic, and quick-thinking human agents involved. It could be a pivotal decision point, a delicate moment where a customer’s loyalty or a sale could be gained or lost, in which a human agent’s careful intervention might make all the difference.
However, the opposite situation is equally true. How many times have you called customer service only to be put on hold while an agent searches for an answer to your question or transfers you to another department entirely? But if customer service agents have tools like IBM Watson at their disposal, they’ll be far better equipped to efficiently resolve customer queries themselves, reducing call times while keeping customers happy, too.
A win/win customer service solution
To some technologists, the idea of A.I. software putting customer service agents out of work is very attractive, and there are areas where tools like IBM Watson can, indeed, perform certain tasks much better than human beings. But for now, having living, breathing agents working together with advanced cognitive computing software provides a far more versatile customer experience solution. Just as futurist Ray Kurzweil envisions the future of A.I., I think that in the world of customer experience, it isn’t going to be us versus them but rather us and them—with A.I. augmenting our own capacities to usher in new heights of human possibility.
That, at least, is the vision of the future I’d like to get back to, even as we rightfully marvel at the ways in which advanced software leaves some of our own capacities in the dust. I bet your customers will prefer that future, too.
Remy Claret has over 15 years of experience in Customer Experience Management. He has worked in a variety of strategic roles in his career including Product Marketing, Sales Engineering and Business Consulting. Prior to joining Product Marketing in 2013, Remy led the Sales Engineering teams for Southern Europe and French speaking Africa. Remy...