Could AI ever fully replace human customer service?

Artificial intelligence is becoming more advanced and many industries are starting to adapt it. When it comes to customer service, does AI pose a threat to human workers? How efficient can it be in handling customer concerns?

artificial intelligence / machine learning / worker replacement
Thinkstock

We’re already seeing the beginnings of artificial intelligence (AI) as applied to customer service, but just how far could it actually go? The idea is beneficial for both customers and employees in the customer service arena; by automating some of the simpler tasks currently done by human employees, customers get what they need faster, companies save money, and employees are freed up to tackle more stimulating, challenging tasks (which can boost morale).

But could AI ever fully replace human-based customer service?

AI in customer service today

Whether you realize it or not, you’re probably already relying on AI—or at least automation—for at least a few of your customer service interactions.

  • Automated ticketing systems. Have you ever submitted a complaint or concern about a piece of software using a contact form? Your request was probably entered into a ticketing system software, which automatically notifies members of the company’s development team to look into and address the problem. There are different ticketing systems for different functions, including software-based tickets and ordinary requests for customer service. Some systems use AI to sort and distribute incoming tickets according to the type of language they use, making it easier for customer service reps to address them.
  • Chatbots. Have you ever had a conversation with a chatbot? They’re nothing new, but they’ve only risen to prominence in the customer service industry—sometimes passing the Turing test—within the past few years. Some companies are even developing realistic human faces to accompany voice chats, so you can read their expressions and feel like you’re getting a more “human” interaction.
  • Contactless phone interactions. We’ve all been forced to go through a step-by-step, automated phone menu to get in contact with a customer service representative, and most of us hate the process. However, phone systems have cleaned up in recent years, with sophisticated voice recognition replacing the clunky press-a-number systems that used to be the pinnacle of technology.

The future

So where could AI in customer service go from here?

  • Ironed wrinkles. The first step is ironing out the wrinkles of systems that already exist, and adding new features that make them even better. The human faces and expressions to accompany existing chatbots is a perfect example of this; we’ll also see improved language recognition, more intuitive ways of organizing and presenting information, and systems that are better at addressing our immediate concerns.
  • Voice biometrics. You’re already engaging with AI by speaking to it, and it’s already understanding the words you’re saying, so what if it could understand how you were saying it? Voice biometrics are being investigated as a possible way to authenticate users over the phone (or in person), such as when trying to gain access to a bank account or credit card information. In addition, machines could be taught, and then self-learn how to read emotional fluctuations in customers engaging with them. They could then soften their tone when necessary, such as if a customer grows angry or impatient.
  • Personalized assistance. AI developers will also spend more time developing methods for personalized interactions. This might mean creating different personality profiles for different customers and circumstances, or automatically adjusting to fit a customer’s specific needs once it gets to know that customer’s preferences. These types of reflexive, adaptable systems will permeate society in many industries and at many levels, and customer service will be no exception.

Limitations

However, there are some hard limitations to what AI can actually do. It seems impossible to develop a machine that can mimic everything about a human being, including vocal imperfections and intonation, not to mention how a human being looks. And no matter what, some people will always prefer human service over service via machine.

In the short term, it’s also necessary for human observers to take note of actions taken by AI algorithms and chatbots; if a chatbot picks up a behavior that customers find unpleasant, for example, it needs to be noted and corrected. In the short term, then, human customer service agents won’t exactly be replaced—they’ll be promoted, transforming the role of customer service agent rather than doing away with it altogether. Most companies will want to have human representatives around to serve customers as a backup plan anyway.

The bottom line

It’s likely that AI will remain a powerful force in the customer service world, and grow to become even more powerful as new technologies develop. It also seems unlikely that AI will ever be able to fully replace human beings—after all, there needs to be someone around to supervise, design and improve these algorithms (at least until the singularity). Still, it’s virtually impossible to accurately predict when, how, or how far technology will develop, so until we see it unfold in the present, it remains an uncertainty.

This article is published as part of the IDG Contributor Network. Want to Join?