top of page
  • Greg Bigos

Triaging Call Centers with AI

Customers hate to wait.

Almost 60% of consumers find that long wait times and lengthy holds are the most frustrating part of a customer experience. Nearly 35% are irritated when they have to repeat themselves during a single service call. And customers are more likely to leave for a competitor if their problem is service-based.

Trying to match staffing with call volumes is a continuous struggle. Employees take vacations, call in sick, or take a day off. No matter how hard call centers try, there's always a mismatch. Too few representatives and wait times increase. Too many customer service agents and operational costs rise. Finding the right balance has been more art than science.

Call centers are starting to view technology as a way to bring more science into the process of managing the customer experience. With advanced technologies such as artificial intelligence (AI), call centers can use their data to develop systems that can reduce costs and improve customer service. In fact, about 20% of call centers are expected to deploy some form of artificial intelligence (AI) by 2022.

Most call centers use some form of interactive voice response (IVR) technology. When callers are asked to enter or say account numbers, they are using IVR. With AI, the initial attempts at call triage can do much more. They can reduce wait times and improve customer service by:

  • Identifying and Routing Calls

  • Prioritizing Emergency Calls

  • Answering Routine Requests

In addition, AI solutions can highlight areas for operational improvement and predict staffing requirements based on historical data. AI can ensure that the critical calls are answered quickly, routine requests are answered seamlessly, and operations flow smoothly.

Identifying and Routing Calls

The goal of AI technology in call centers is to put callers in touch with the right person to answer their questions as quickly as possible. Existing IVR systems have spent decades trying to identify and route calls as quickly as possible. However, many solutions are binary, meaning the customer may have to go through a number of interactions before their call can be answered.

We've all experienced the automated narrowing process. "If you're looking for our office hours, press 1. If you're wanting to make an appointment, press 3." AI can interact with the caller using natural language processing (NLP) to determine what the caller needs. The flexibility allows customers to explain exactly why they are calling instead of struggling to select an option that fits. The more an agent knows about the caller and the call, the better they can address the problem.

For example, AI can learn to detect emotion through the language we use. Suppose the technology detects a growing irritation from the caller as it tries to narrow the caller's request. At that point, the call could be routed to a human even if the AI solution could respond. To deflect caller frustration, the solution routes the call to a human for immediate attention. At the same time, AI can display all the information it has collected and highlight the increasing agitation of the caller, ensuring the agent is better prepared to talk with the customer.

Prioritizing Emergency Calls

Most callers believe their call should be given priority. It may not be an emergency, but it's crucial that they get their answer ASAP. Unfortunately, their call may not be a true emergency. How many call centers have received an "emergency" call only to discover it wasn't even close? Call centers can identify emergency calls with AI and route them to the appropriate resource before the caller even speaks with an agent.

Using company-supplied criteria for identifying emergency conditions, AI can learn how to intuit if the caller meets the criteria. Over time, the technology compiles different terminology that is used to express the same situation, which increases its effectiveness. It's possible for AI to learn where to route the call based on the emergency.

For example, a client calls to say the network is down. The solution collects some basic information before passing the call to the connectivity team to address. Rather than sending the call to customer support, AI can learn which groups are best at responding to specific customer questions and route the call directly instead of through a Level One pool of support personnel.

Answering Routine Questions

How many times a day do agents answer the same questions? Five, ten, or twenty? It doesn't take long for a hint of irritation to creep into their voice. Unlike humans, AI doesn't care how many times a question has been answered. It will respond to the question in the same way, no matter how many times it has been asked.

Using chatbots, for example, AI can direct the caller to a self-service page for an answer or offer the solution as part of a chat conversation. Minimizing the number of routine calls call center staff must answer frees them to spend more time on calls that require complex solutions.

Improving Operations

AI can help call centers balance staffing with workload. Using historical data, it can determine what time of day or day of the week an organization receives its peak period. It can make tweaks to the system to adjust to changing patterns. Because AI is processing more data more efficiently, it is able to detect patterns that humans may not see.

Want to learn more, or meet virtually to discuss your call center needs? Contact us today to connect.

49 views0 comments

Recent Posts

See All
bottom of page