Artificial intelligence (AI) is one of the hottest topics in call centers today. Strategic debates regarding its application for customer service, the potential for job disruption and the ethical issues surrounding human work displacement are taking place in boardrooms around the globe. But while AI is still in its early days—primarily in discussion and planning stages—the hype storm surrounding it already has contributed to inflated expectations that far exceed reality, according to Gartner’s Hype Cycle for Emerging Technologies.
AI is the most recent buzzword that technology companies and marketing firms are embracing and infusing into their content and sales pitches as the must-have innovation. And yet, the call center industry is no stranger to this type of frenzied buildup for technology that promises to solve service issues and lower labor costs by replacing agents with software or applications. In the late 1990s, online support was the highly touted successor to call centers. A decade ago, marketers latched onto social media as the shiny, new replacement to call centers. Well, we’re still here!
We’re not buying the hyperbolic sales pitches from technology companies foretelling the end of human interaction in call centers. It is simply irresponsible rhetoric. We don’t believe that R2D2 will replace the human agent anytime soon.
Studies have shown that the average consumer would rather interact with a live agent than a chatbot. A recent Forrester survey found that, if offered a choice, 83% of consumers said they would prefer to speak to a human since human agents better understand their needs (78%) and can address multiple questions at once (57%). The top three problems that consumers reported with bots were that they could not deal with complex requests, deliver personalized offers as well as humans, or understand human emotions.
That being said, it would be unrealistic to dismiss the potential impact of automation in the call center, particularly when AI is combined with robotic process automation (RPA). We believe that call center agents will continue to provide a critical touchpoint in the customer journey and that the technology, over time, will be used to augment human performance in the call center to deliver more efficient, effective and value-added experiences.
Ask six people to define AI, and you’re likely to get six very different answers. Many of the claims, exaggerations and fears around AI stem from differing views on what it is and how it can be applied in the call center.
To understand how AI and RPA can work with human agents to improve service delivery, first let us clarify what we mean when referring to these technologies used in a call center context. The following is what we believe them to be:
RPA can be “unattended” or “attended,” depending on its use.
RPA is comprised of “dumb robots” that require rule-based processes and a set of instructions, after which they will perform the same tasks over and over in the same way, consistently and accurately.
AI, on the other hand, is self-learning and designed to simulate some of the behaviors associated with human intelligence in order to solve complex problems.
As with RPA, there are two types of AI—narrow (or “weak”) AI and general purpose AI.
Call centers that have large volumes of repetitive processes—such as data entry and migration, invoice processing, loan processing, claims processing, etc.–can increase productivity and efficiency by assigning those tasks to RPA robots. A more significant ROI of RPA, though, may be the impact on staff engagement. One of the top complaints agents have about their job, and a critical factor in the decision to leave, is that they find the work to be too repetitive and boring. Removing tedious tasks frees agents to focus on the more valuable and interesting aspects of the work that require creativity, decision making and interacting with customers.
Combining AI with RPA can provide call centers with an ideal scenario for using automation with machine analysis. For instance:
Despite consumer preference for interacting with human agents, chatbots can be effective in specific situations and can free agents to deal with the more pressing customer issues. A recent study, “2018 State of Chatbots,” revealed that consumers saw the potential benefits of interacting with chatbots when they needed 24-hour service (64%), instant responses (55%) or answers to simple questions (55%).
While the benefits of augmenting human performance with intelligent robots seem appealing, machine learning is still at least two to five years away from mainstream adoption, according to Gartner. Some AI subsets like cognitive computing are up to 10 years away.
RPA adoption, however, has already begun but with limited success. Ernst & Young reports that it has seen 30% to 50% of initial RPA projects fail—but not because of the technology. RPA requires long-term planning, and the process is much more complicated and resource-intensive than many executives had assumed (or technology vendors had acknowledged).
In fact, many early adopters have been putting their RPA projects on hold while they reconsider how bots can be applied as part of a coordinated system, rethink tasks and workflows, put more resources in place to manage the bots, and train employees how to use bots to solve problems. According to McKinsey & Company, the journey from technical automation potential to full adoption is likely to take decades.