In 2019, artificial intelligence (AI) is taking the world by storm. The optimism around this emerging technology spurs businesses around the world to pour investments into AI, with the hope that it will transform their efficiency and profitability. In February, President Trump signed an executive order calling for federal agencies to invest in AI to grow the economy and achieve greater national security.
Perhaps unintentionally, the executive order epitomizes a key trend in AI today: widespread readiness to invest with little insight into budgets or timelines for adoption. Although plenty of companies are eager to harness the power of AI, few are able to accurately gauge its long-term impact.
Customer service represents one of the most appealing—and riskiest—applications of AI. More and more companies are augmenting their customer service with automation software, chatbots and call center scripts. These AI-powered tools promise to filter queries quickly and efficiently, at a fraction of the cost. After all, chatbots don’t need wages like the human customer service reps do. The average annual salary of a customer service representative is $35,800, compared to just $3,000 to $5,000 for a Facebook Messenger chatbot.
But automation doesn’t necessarily save you money in the long run—and testing the waters of AI adoption could expose your company’s reputation to serious risks. Here’s why it’s still worth investing in customer service with a human touch.
Patchy data leads to AI difficulties
A company’s customer service directly impacts its overall reputation, which in turn determines its bottom line. Before rolling out an AI solution, companies must weigh the benefits of faster service against the risk that AI could deliver a subpar experience for customers.
In an ideal world, AI tools can help companies serve more customers, especially outside of normal business hours or during periods of high demand. For example, when customers call a company’s support line, many are forced to endure hold music while they wait for the next customer service representative to pick up the phone.
By contrast, an AI chatbot isn’t limited to helping one customer at a time or working during business hours. Instead, it could help an unlimited number of customers instantly, at any time of day or night.
The problem is, a lot of companies are working with bad data, whether it’s incomplete information or too many data sets to meaningfully manage. Poor data management can amplify a bad customer experience and harm businesses.
In fact, 95 percent of respondents in a recent Experian survey said that bad data negatively affects their business; it costs $3 trillion a year in the U.S. alone, according to the Harvard Business Review.
Bad customer service will cost you
Additionally, there’s a big difference between using AI to improve overall customer experience compared to using it in real-time customer service interactions. While AI might help companies manage customer experience at a high level, the reality is that the most memorable customer service experiences occur when a customer is unhappy.
Telestra, an Australian telecom company, learned this the hard way when it launched a virtual assistant chatbot nicknamed “Codi.” Codi struggled to interpret customers’ questions and connect them with the information they needed.
Worst of all, when customers requested a human agent, the bot promised to connect them, only to promptly get stuck in a loop. One customer asked for a human agent 15 times during a single chat.
Frustrated customers took to social media.
As customers hop on social media, it prompts local news outlets to cover the disaster. American publications later picked up the story, piling on the negative press. Instead of improving customer service with AI, Telstra ended up with angry customers and a significant reputation issue.
The risks for an unhelpful AI interaction remain incredibly high. The 2017 American Express Customer Service Barometer found that more than half of Americans have abandoned a transaction after receiving bad service, and one-third would consider switching companies after just one negative service experience.
Businesses can use AI to leverage data and improve customer experience.
AI’s ability to provide direct customer assistance remains unproven. Younger generations, especially, are more skeptical and less likely to wait patiently for service. Until the technology improves, most companies should continue to trust human representatives to handle sensitive interactions with unhappy customers.
Human customer service is still a worthwhile investment
Maintaining well-trained, human customer service staff—especially one that’s available 24/7—requires a significant investment. But that investment delivers excellent returns.
According to a recent PwC report, an astonishing 82 percent of U.S. consumers want more human interaction in customer services. Customers are also willing to spend up to 16 percent more on products and services if they receive great customer service, resulting in long-term customer loyalty.
Costco, a market leader, often lauded for its great customer service, it also pays its employees higher wages and gives them more benefits than the retail industry average. Investing more in customer service yields higher customer loyalty and satisfaction, creating a positive feedback loop for growth.
Of, course, this doesn’t mean companies should ignore AI, but the best approach is to combine the best of both worlds. Businesses should use AI to supplement, rather than replace, human customer service representatives—and do so slowly and carefully.
A well-designed chatbot can answer low-level, FAQ-type questions from customers.
Using a chatbot for lower-lever questions frees up human customer service representatives to focus on more complex issues. Chatbots of the future may even be able to compose responses indistinguishable from real humans, and with a literary flair, to boot.
Slack provides a great example of this hybrid approach.
Every Slack account is equipped with a channel called “Slackbot,” where users can type simple questions about how to navigate their accounts. Its customer service channel, however, is operated 24/7 by real people who aren’t allowed to use scripts.
If Slackbot isn’t able to solve a problem, customers still have access to human representatives who never circulate through automated responses. This commitment to customer experience has played a pivotal role in Slack’s explosive growth.
Automation, AI and data analytics are all an integral part of customer service—but they’re not sufficient on their own as of yet. Today, investing in great customer service staff will still pay off.