AI Is Changing CX, but Human Intelligence Still Matters
Artificial intelligence is radically reshaping the modern customer experience, powering sophisticated support automation systems that enable businesses to deliver faster, better customer service at scale as their operations grow. Unlike their human counterparts, AI-powered virtual agents have the ability to provide uniform quality of service for customer engagements, 24 hours a day, with little or no downtime. However, while some have argued that these advantages will quickly make human support agents obsolete, they fail to account for the very real limitations of AI technology. Ultimately, modern artificial intelligence is only as “smart” as the data and input we give it. For the foreseeable future, human training, oversight, feedback, and — crucially — expertise will have a vital role to play in automated support systems.
Support Automation on The Rise
According to a report issued in January by global research and advisory firm Gartner, enterprise-level AI implementation has grown by 270 percent in the past four years, and 37 percent in 2018 alone. A more recent Gartner survey found that nearly 40 percent of service leaders are either piloting or actively using bots and virtual customer assistants in their contact centers.
The value of AI in customer service is hard to dispute. Automated support systems allow companies to efficiently solve simple issues at scale, as AI-powered virtual agents have been shown to drastically reduce customer wait times. However, per Gartner CIO Research group Chief Researcher Chris Howard, “we still remain far from general AI that can wholly take over complex tasks.” And as any service professional will tell you, human interaction in the world of customer support is far from a simple task.
Complex Problems Require Human Intervention
Virtual agents excel at helping customers solve simple problems quickly and at scale. Given a sufficient content library of information about a company, its products, and solutions to common customer complaints, they can easily retrieve appropriate responses to straightforward inquiries. For example, “The nearest store opens at 8 a.m.;” “Refunds are accepted until 30 days after purchase.;” “Just restart your phone to get it working again.” Virtual agents can instantly deliver these simple solutions to customers when they need information fast.
However, it’s rarely that simple. Without human assistance, an AI-powered virtual agent is only as good as the data it has to interpret questions and deliver answers. And, if the virtual agent can’t understand the customer’s intent, it doesn’t matter if you have a great content library for the virtual agent to pull from — taking into account cultural and regional differences, company jargon and slang, the way customers ask questions is incredibly varied and difficult to decipher. Businesses need to build vast databases to give their AI enough information to tackle the wide variety of customer engagements and support inquiries it may encounter, because a virtual agent is merely an imperfect facsimile of a resource we’ve been using to solve our problems for thousands of years: human expertise.
A company’s virtual agent generally shares many traits in common with a human who is an expert in that company’s product or field. Both possess vast amounts of useful information. Both are able to share that information with customers who need it. However, only the human expert possesses our unique ability to interpret data, synthesize information, and devise creative solutions and empathetic responses to problems. Like all human interaction, customer engagements are full of “edge cases,” inquiries and requests that are simply impossible to predict and catalogue in a pre-written content library. These edge cases call for good, old-fashioned human expertise.
Frustrated Customers Require More
Content libraries enable virtual agents to provide customers with the information they need, but that’s not enough to deliver a complete, optimal customer experience. Virtual agents also need to understand intent — e.g. what the customer is actually asking — in addition to robust language libraries to present information in a way that is both easy to understand, and aligns with a brand’s values and tone of voice.
When it comes to customer support in particular, frustrated or unhappy customers don’t just need solutions — they need empathy and respect. Yet once again, even the most extensive, thorough, meticulously-constructed library is insufficient to address the full spectrum of customer emotions that a virtual agent may encounter. As language libraries and intents are built up over time, virtual agents can certainly learn to better respond to unhappy customers with empathy. However, a virtual agent doesn’t inherently have this ability, and can easily make the customer even more upset, exacerbating an already irksome experience.
Many consumers still avoid interacting with virtual agents for these very reasons. In a recent survey of call center agents, 85 percent of survey respondents agreed that customer web-to-phone escalation often involved a frustrated customer requiring a live agent to deal with complex issues. 82 percent of survey respondents also felt that customers sought the reassurance that human interaction brings to the conversation. The facts speak for themselves. Virtual agents may be unbeatable when it comes to tackling simple problems at scale, but unless they have the data necessary to handle a particularly complex question or an especially emotional customer, they still need human help. This is where human oversight becomes imperative.
Combining AI Solutions with Human Expertise
Today, the best customer experience management is being performed by companies that combine AI- and human-powered services and customer support. These companies use virtual agents to filter out repetitive, straightforward customer engagements at scale—allowing them to provide 24-hour service at minimal cost. However, they also ensure that virtual agents are ready and able to escalate issues to their human counterparts when customer needs exceed the virtual agent’s capabilities, ensuring a more positive experience overall. It will take time for artificial intelligence to advance to a stage where its able to tackle complex problems with the sensitivity and respect that customers demand. Until then, AI will continue to have a big role to play in customer support, and so will human intelligence.