Artificial Intelligence in the Call Center

3 Responses to Using AI to Serve Customers

By Peter Lyle DeHaan, Ph.D.

Predictions about the emergence of artificial intelligence (AI) have been with us for decades. But until recently they only showed up in science fiction books and movies—usually with dire results. Such is the basis for good fiction.

Yet in recent months, advances in artificial intelligence have surged forward, reaching into every industry, including the call center and customer service sphere. With AI, just as with any technological advancement, there are three responses.

Ignore It and Maybe It Will Go Away

The first reaction, which is really a nonreaction, is to dismiss it. Maybe you’re already sick of the hype or maybe you’re not aware of it. Yet assuming a computer algorithm has no place in your call center is not a wise conclusion to make.

The risk of this approach is getting left behind. You will find—likely in short order—your call center operation and your company competing with others who have thoughtfully integrated artificial intelligence into their operation.

They will serve customers in a way you cannot and save money you’re not able to.

Gung Ho Adoption

The second response is the opposite. It’s to go full speed ahead in adopting artificial intelligence technology for the call center. Yet this is also fraught with peril.

The news is filled with artificial intelligence going awry. In recent months, companies have been publicly embarrassed and their stock has taken a hit, not because of human error (at least not directly) but because of computer error. These occurred from AI applications running unchecked and without restraint.

If you’ve ever used text chat to submit a customer service request, you’ve likely interacted with a chat bot, which is an artificial intelligence application. In my experience they’re unlikely to solve my problem, but usually they collect some preliminary information and route me to a real person who can help.

Yet just recently, a chat bot took me down the wrong path, leaving me with two unacceptable options: agree that the chatbot had solved my problem or pay to upgrade my service. End of discussion. But it wouldn’t allow me to start a new chat session until I concluded the first one by picking either of its two unsatisfactory answers.

I also think artificial intelligence was involved in a recent near-miss with an email support effort. I had submitted a service ticket, but a couple hours later I figured out the solution on my own. I sent a follow up email to cancel the ticket. The response told me how to cancel my service with the company. This may have been a human error by an agent who scanned and didn’t read my email, but I suspect it was artificial intelligence which responded wrongly to the word cancel. Fortunately, the AI bot didn’t take the initiative to close my account.

Imagine seeing these examples extended to telephone calls at your call center. Yet it’s already happening.

I recently read a report of artificial intelligence telling human agents how the solve customer problems and what to say. The AI then grades the agent on compliance, penalizing agents who use common sense to override the AI’s bad guidance.

Then the common excuse of “I was just following orders,” becomes “I was just doing what the computer told me to.” May it never be.

Cautious Implementation

The third response—the one I recommend—is a balanced perspective. Investigate the use of artificial intelligence in your call center operation. Make an informed decision as to how to best use it. The wise application is to implement artificial intelligence to better serve customers. Don’t pursue AI merely to save money, even though this should emerge as an expected outcome.

Seek ways where artificial intelligence can make your agents’ jobs easier. Look for ways where AI can help your human staff better serve your human customers. A guiding principle in this is to keep AI in an advisory capacity. Give your agents final say. They should be able to control the AI, not have the AI control them.

As you appropriately implement artificial intelligence in your call center, the goal should be to offer better customer service, improve response times, and lower payroll costs. But don’t look for AI to replace your staff anytime soon. And my advice is to resist the urge to blindly implement AI, lest you end up with a public relations nightmare, lost business, and a decrease in new customer acquisitions, all through AI run amok.

A good baseline requirement to guide your use of artificial intelligence in the call center is to empower your agents to control it, not let AI replace the common sense and empathetic problem-solving ability of real people.

Peter Lyle DeHaan is the publisher and editor-in-chief of Connections Magazine. He’s a passionate wordsmith whose goal is to change the world one word at a time. Learn about his books and read more of his articles at  Peter Lyle DeHaan.