Beyond Chatbots: How Artificial Intelligence Can Humanize the Customer Experience

The rise of artificial intelligence has led many people to believe that machines will take over many of the job functions in the marketing industry. However, when it comes to customer experience, the best results will occur when machines and humans work together, writes Dan Gingiss, VP of Marketing at Persado.

When people think of artificial intelligence (AI) and its technological cousin, machine learning, it often conjures up images of a robot taking over the job of a human. Indeed, “the robots are coming!” is a familiar refrain suggesting the human race will soon be invaded. If we can teach the machines to be as smart as humans, consume mountains of data in seconds, and continuously learn more at warp speed, isn’t the useful life of the human brain quickly coming to an end?

THE NEW RULES OF (MULTI-CHANNEL) CUSTOMER ENGAGEMENT: HANDBOOK FOR GROWTH MARKETERS

Winning CX will come from Brands who can balance relevance, consistency and convenience to drive engagement. The kind of engagement that drives optimal customer lifetime value and real business impact.

Download

The answer: Absolutely not.

Also Read: 5 Customer Experience Tips from Experts on MTA

Anyone who watched IBM’s Watson destroy its human competition on Jeopardy! knows that computers are getting really sophisticated. Professor Toby Walsh of the University of New South Wales said that AI will catch up to human intelligence by 2062 — learning unique human traits such as adaptability, creativity and emotional intelligence. But that’s still decades away.

Today, and in the foreseeable future, the best results come from humans and machines working together. This applies to all aspects of business, even those parts of the customer experience that feel emotional and are inherently relationship based. In fact, when applied strategically, artificial intelligence can enhance a customer’s interaction with a company and, in some cases, even humanize it.

Data-driven marketing that’s inspired by humans, validated by AI   

To start, artificial intelligence is becoming more prevalent in marketing at the top of the funnel. This isn’t surprising, as a lot of content across industries like finance (such as earnings reviews) and sports (such as game summaries) has been generated by machines for a while now. In fact, machine-generated content has gotten so good that most people have a hard time distinguishing it from human-generated content.

For example, compare these two real email subject lines from an online pharmacy:

1. ALERT | We noticed you may need to order your refills.
2. Don’t wait! Order your refills with one click.
    

Do you know which subject line was written by a human and which one was written by a computer? Are you sure? In a recent audience of about 75 marketers at Dreamforce, about 90% guessed incorrectly. (See the end of this article for the answer.)

But AI in marketing requires human intervention. The machine-generated subject line above still required human input.

Burger King recently launched an ad campaign that was claimed to be “created by artificial intelligence.” The headline, describing a new chicken product, was “Tastes Like Bird.” The company even issued a press release saying that it was “moving from the traditional Agency of Record to an Agency of Robots.”

The company admitted that the ad was a stunt, but it was backed by a real message

“Artificial intelligence is not a substitute for a great creative idea coming from a real person,” Burger King’s global head of brand marketing told AdAge ($).

While the ad succeeded in getting a lot of laughs and media attention, the resulting sentiment that AI is not a substitute for human creativity is spot-on. What machines can help marketers do is evaluate thousands of combinations of messages to determine which ones will be most effective at any given time. Just as Watson could conjure up far more potential Jeopardy! answers than its human opponents, a machine can generate far more messaging combinations than a human copywriter. And it can do it at scale, as the computing capacity is no match for a human brain.

Machines can also track decades of historical campaigns and analyze that data to provide insights and trends. After testing, a machine can identify why a certain message worked, not just that it outperformed a control message. For example, the presence (or absence) of personalization can be the most significant driver of consumer responsiveness.

(Of course, asking a machine to replace human creativity entirely results in unusable headlines such as “Tastes Like Chicken.”)

A new era of bots that support human interaction

Outsourcing customer service -- arguably the most human resource-intensive part of the customer experience -- to a “robot” seems like it would be bound to fail. But surprisingly, there are many strategic ways to apply AI to better support the human beings on the front lines with customers.

One major enabler is the proliferation of messaging apps. Three of the five largest social media platforms in the world — WhatsApp, Facebook Messenger, and WeChat — are messaging apps, each with more than a billion active users according to Quartz. If those apps were countries, they would each be more than three times the population size of the United States.

Joshua March, CEO of digital customer care platform Conversocial, said two years ago that he believed messaging apps were “the future of customer service,” and he was right. His reasoning was that messaging apps have two main advantages over “traditional” social media: they’re private (which brands love because it keeps complaints out of the public eye) and they’re persistent (which customers love because they don’t have to repeat their problem or even their account information to multiple agents).

This shift has forced companies to adapt quickly, because the move from public to private channels has not decreased consumers’ extremely high expectations of a seamless customer experience. As a result, many companies have turned to AI.

Customer service “bots” have been around for a long time —in the form of Interactive Voice Response (IVR) systems on the phone and even some very basic automation on website click-to-chat services. The digitized version of this will most certainly need to be a better, more personalized customer experience, and that’s where bots are still lagging behind. Forrester predicts that in 2019, “Human resistance against ineffective chatbots is on the way, and a groundswell of jaded customers will crowdsource tips for end runs around chatty chatbots.”

So how can chatbots be most useful? The best place to start is routine customer service inquiries, of which there are many. The reason why that IVR on your credit card company’s toll-free number starts with “To hear your balance” is that millions of people call in for just that reason every year. If messaging bots can provide the same service with frequent or repetitive inquiries, it should result in a faster, easier customer experience for the consumer and a more efficient process for the company, which can then focus its agents on more complex inquiries that need human interaction.

The other place where bots can be of value is in support of the agent. Intelligence Augmentation, or IA, is sort of the reverse of AI. Instead of the bot taking over, it assists the agent in delivering a human response to the customer. Imagine an agent having a Watson-like companion constantly at his or her side which can look up customer information and answers to millions of questions in a fraction of the time the agent can? Then the agent is still there to provide actual human engagement.

This will require new training and processes to teach humans how to work with their machine counterparts. “Smart companies know that AI investments must include concurrent investments in education on the impact of AI in their business and employees and how expectations, skill sets, and experiences will evolve,” writes Forrester.

But too many companies will invariably look at bots as a shortcut, a cost-saving opportunity that will result in “IVR hell” again, just on another channel. Instead, bot experiences should be designed to incorporate human interaction at key moments -- and that handoff needs to be both simple and seamless.

Also Read: Does Customer Service Improve Customer Experience?

When humans and AI work together

So what works best? When the machine and the human work together. “Strong CX will depend on getting machines and humans to work collaboratively,” writes Forrester. Self-service is great, and many customers prefer it, but when there’s a problem, they don’t want to talk to a machine. As for marketing communications, human inputs -- such as branding requirements, suggested emotions and preferences of leadership -- can provide the initial inspiration and strategy, helping steer the computer in the right direction from the start.

The robots may eventually take over, but for now, the human touch reigns supreme.

(Incidentally, the answer to the email subject line challenge above is that #1 was created by a machine, and #2 was created by a human. And the machine-generated subject line resulted in a 23% higher open rate and a 73% higher click-through rate.)