Use AI to Empower Agents, Not Replace Them
There is a new adage in the business world: You won’t be replaced by AI, you’ll be replaced by someone using AI. For contact centers, that certainly rings true. Over the past several months, the onslaught of doomsday headlines surrounding artificial intelligence (AI) taking over the human race have become de rigueur.
Across several industries, companies are racing to adopt this buzzworthy technology but many organizations could be losing sight of the forest for the trees. While contact centers were one of the first to adopt AI applications like chatbots to streamline interactions and triage customer service requests, in their haste to adopt AI, the focus has been more tactical than strategic.
Stand-Alone or Omnichannel: Meeting Customers Where They Are
Through the 1990s, contact centers focused entirely on how to better handle phone calls, and as a result, customer satisfaction numbers improved. When chat and email were first introduced, “meeting customers where they are” soon became a catchphrase in the industry. Consumers loved that they could now contact a company on multiple channels. Then, in the 2000s, customer satisfaction numbers started to fall. Could the addition of email and chat channels be responsible for the decline?
The truth is, even adding a single additional channel significantly increases the complexity, and every new channel multiplies this complexity. So when chat and email channels were added in the early days, the immediate side effect was a drop in customer satisfaction. Customers found that although they could contact companies via chat, the information they received was siloed and potentially incorrect. In my own experience, when I contacted a company to resolve a cross-border billing issue, the call center had problems resolving the issue, but their Twitter presence advertised that they could deal with complex problems. However, the result of my Twitter conversation was the suggestion that I contact the call center.
And therein lies the problem. When channels are added without giving thought to how they are handled internally, they create data and organization silos that do not share information or processes with each other. Unfortunately, to this day, many contact centers aren’t investing wisely in omnichannel solutions and choose to invest in short-term solutions instead.
Rip & Replace: Why It’s a Mistake
AI in a contact center can be quite effective and is routinely used for contact routing and allocation, sentiment analysis, and agent guidance, but it doesn’t go deep enough. Which is why the frontline agent job isn’t going anywhere, at least not just yet.
Many companies have been in the headlines recently, announcing job cuts in the thousands at contact centers. The impact of this “rip and replace” approach, where companies are rushing to eliminate frontline jobs, has the potential to exacerbate or hide problems caused by siloed customer communication channels. Before ripping out and replacing current tech, you must evaluate if its replacement will provide flexibility without forcing you into a pigeonhole. Ideally, it lets you build on, integrate with, and utilize the technology and equipment you already own.
Challenges with Yet Another Form of Self Service
Too many companies are rushing to implement AI in a vacuum without considering how it may impact the agent experience or customer service. The problem with shortsighted implementation is that not all of your organization’s requirements can be addressed by a single AI application. In addition, AI can add increased desktop complexity to the unified agent desktop view, as it can introduce new siloed channels. Plus, implementation of poor self-service features could lead to customer dissatisfaction and duplicated costs (self-serve plus assisted service), and will increase both customer and agent frustration (customers are upset that their time was wasted in self-serve and could take it out on the agent).
At its best, AI can serve to process complex information. It can parse through speech to determine the underlying request and properly categorize leads. When agents are asked to provide specific information to a customer, they need to find the information requested and ensure its accuracy. Specifically, in the “agent assist” case, AI doesn’t have to be 100 percent correct, as long as it provides the right options. Agents are important information brokers in this situation, making the AI easier to implement and less risky to operate.
Using AI to Augment Agent Capabilities
Adding AI without proper planning will increase agent desktop complexity and will impact the bottom line with increased calls, more errors, more agent burnout, and ultimately customer dissatisfaction.
It is important to identify mundane tasks where AI makes sense and the kinds of tasks that are better performed by humans. Generative AI in chatbots is certainly having a moment today, but conversational AI is great at exactly that—having a conversation or processing information. Providing actual service still relies on proper operationalization into the business to be able to supply authentic answers, and it would be disastrous if left unsupervised with AI.
Agents need to be able to process a lot of information quickly and correctly, and with increasing complexity as a result of improved self-service. Agent training either has to be supercharged to deal with “long tail” issues, or agents need tools to assist them in doing their jobs effectively. As organizations spend time reducing costs through AI-powered self-service, contacts routed to agents will have more complex issues that will require a human touch to resolve.
We should view AI as an augmentation to the agent, or as an augmentation to self-service. People who want to use self-service for routine inquiries will benefit from credible and well-operationalized AI applications. But agents who use AI will be able to provide better, more consistent service. The brand experience is best served by using AI visibly to improve self-service options, and invisibly to improve agent abilities. When combined, both will improve brand appeal by giving the customer choice along with more accurate and faster service.
Rob McDougall is president and CEO of Upstream Works Software, which provides omnichannel conact center software to increase customer engagement and agent success. McDougall’s passion for making the agent experience better is evident with the company’s “agent first” approach.