Hello, Hal! Artificial Intelligence In The Contact Center

AI for customer service has enormous potential, but it’s important to maintain a realistic perspective.

While HAL was able to converse naturally with people on any imaginable topic, the conversational user experience with AI still isn’t there and probably won’t be anytime soon.”

Paul Stockford

Paul Stockford is Chief Analyst at Saddletree Research, which specializes in contact centers & customer service.

In 1968, a movie called “2001: A Space Odyssey,” billed as “An epic drama of adventure and exploration” was released. Produced and directed by Stanley Kubrick, one of the most influential directors in cinematic history, “2001” told the story of a space voyage to Jupiter on the United States spacecraft Discovery One. On board Discovery One are two pilots, three scientists and the ship’s computer, HAL 9000 (Heuristically programmed ALgorithmic computer), called Hal by the crew. Hal controls most of the ship’s operations.
Powered by artificial intelligence (AI), Hal is considered to be a dependable member of the crew and engages in conversational English with other crew members. He plays chess with one of the pilots and possesses cognitive capabilities equal to or better than his human counterparts. As the film progresses, Hal becomes the movie’s main antagonist and (Spoiler alert!) things don’t end well.

While “2001: A Space Odyssey” is recognized today as one of the most influential films ever made, what isn’t as commonly known is that Kubrick also had a hand in making a 2001 movie called “A.I. Artificial Intelligence.” In the early 1990s, Kubrick began working on a film version of a short story called “Supertoys Last All Summer Long.” It was a story about a robot that resembles and behaves like a real child and his efforts to become a “real boy” a la Pinocchio. Although Kubrick died in 1999, before the movie was finished, Steven Spielberg picked up the ball and ran with it. The film was released in June 2001 with Kubrick receiving a posthumous production credit.

Although robots that think, act and speak like humans are still the stuff of fantasy, you’d never know it today by the avalanche of press releases, webinars, marketing materials, conference sessions and briefings on the topic of AI in the contact center. Don’t get me wrong—there is a place for AI in the contact center and it will continue to grow in the future, but don’t expect a clone of HAL to replace any of your agents anytime soon. What we’re hearing about today for the most part is not true AI, it is machine learning technology. The difference? AI is defined as an intelligent agent that perceives its environment and takes actions that maximizes its chance of success at some goal. AI mimics cognitive functions that humans associate with other human minds.
Machine learning, on the other hand, gives computers the ability to learn without being explicitly programmed. Although it could be argued that machine learning is a subset of AI, machine learning still relies on algorithms and pattern recognition that it can learn from in order to make data-driven predictions or decisions. In other words, machine learning functions in operational rather than in cognitive terms.

Still, there is no doubt that machine learning will have an impact on the contact center in the years to come. Forward- thinking companies like Cisco are already making moves to firmly plant their corporate feet in the world of machine learning and AI in the future.

“We believe that AI will change the way all software is developed. We are moving quickly to harness the power of cognitive software to make our collaboration and customer care software deliver more value with less effort,” stated Tod Famous, senior director of product management, Cisco Customer Care.

While HAL was able to converse naturally with people on any imaginable topic, the conversational user experience with AI still isn’t there and probably won’t be anytime soon. Instead, companies like MindMeld, acquired by Cisco this past May, are focusing their efforts on AI-based conversational platforms that enable people to communicate naturally within one domain to complete a specific task. Still, the day when someone contacts customer service and reaches a bot that says, “How can I help you?” is still too far away to even predict.

“The most successful natural language dialog technology deployments we see in the market today are designed to support a narrow set of customer care use cases,” Cisco’s Famous continued. “The range and depth of the automation will expand quickly in the future.”

In the “2001” movie, Hal states that he is “foolproof and incapable of error.” Unfortunately, that won’t be the case with AI in the contact center. According to Cisco’s Famous, “We also believe that even the best automation solutions will need customer care teams to handle the exceptions. It’s like the ‘zero-out’ option on an IVR solution. We strongly believe that the best way to support customer success with automation is to offer a smooth escalation experience when the bot isn’t capable of responding to their request.”

Given the allure and mystery of AI, it’s easy to see why the industry has been captivated by it. It’s important, however, to maintain a realistic perspective regarding the potential of AI in a customer service role. The contact center industry has a history of getting ahead of itself, and sometimes ahead of its customers. Granted, the next generation of customers—you know, the ones who look at their phone for the answer to everything—may be comfortable interacting with a robotic customer service agent via text, but there is still a large market who prefers human interaction today and likely will for many years to come.

The development of AI for customer service is in its earliest iteration and while it has enormous potential, there are still a lot of great customer service solutions available today. These proven solutions are not obsolete and won’t be for many years to come. While it’s fun to watch and talk about AI in the contact center, follow Cisco’s outlook and keep it real. Don’t let your head get too far into the clouds. I’m sure HAL would tell you the same thing.

Paul Stockford is Chief Analyst at Saddletree Research, which specializes in contact centers & customer service.

– Reprinted with permission from Contact Center Pipeline,

RSSComments (0)

Trackback URL

Leave a Reply

You must be logged in to post a comment.