Two proposals for call routing in a contact center

One of the eternal problems of a contact center is to achieve that the call finishes in the right agent’s position. This is doubly true in a mobile world that requires almost real-time conversational support.

Let’s add to the combo the fact that most customers seeking support in a contact center are there because they were unable to find an answer to their inquiry previously using a FAQ list, a community or a self-service system. In this situation, people expect an answer in a few minutes.

In addition, the ability to connect with the company in an easy, efficient and ideally cheerful way is becoming more and more a distinctive factor. The customer experience is the result of the links, and for humans the experience gained from the more recent interaction tends to have a greater influence in relation to the previous ones.

Consequently, a positive customer experience matters, not only at the time of the sale, but even more so in situations that require the active help of the company that sold the product or service. That is why it is important to allow customers to get the solution to a problem in the easiest and humanly possible way.

The problem for some contact centers is that when receiving a call or incident, the processes are designed to ask the customer where to begin answering. It prompts the user to qualify and to route it to the correct queue. And this is a task for which he or she is obviously poorly prepared. Most of the time this leads to an unnecessary workload, caused by re-routing the incident, delays in handoffs, in the resolution, and ultimately frustration at both ends: the end user and the service agent.

Why not use the incident itself to automatically identify where to direct the query? There are two basic possibilities for achieving this intelligent routing. We could define one of them as the classic approach and let’s call the other the “text mining” approach.

The classic approach

Automatic routing can be easily accomplished using metadata provided by the IVR application as part of the incident report. Incidents are labeled based on this metadata. Labels are used to direct the incident to an appropriate queue or agent. The configuration labels and routes without explicit user intervention, who is already suffering enough because there is a problem he or she could not solve.

This is a proven approach that requires an elaborate metadata design, its in-application design, and the ability to create sets of rules over these metadata in the service back-end.

With sufficient data and analytical tools, this approach can also be used to provide preventive support, something that significantly improves the customer experience because it can even avoid the negative feeling in the first instance.

However, this approach finds its limits in the ability to implement an application so that sufficient metadata situations can be identified for appropriate labeling.

The text mining approach

With the increasing number of software solutions running in the cloud, improvements in artificial intelligence and natural language recognition applications, it is possible to use additional information: the description that the user submits as part of the incident report.

Although users do not always provide accurate information, it is regularly sufficient to increase and improve labeling for better routing and, ultimately, faster resolution.

Given that in many systems, first-line agents can resolve incidents themselves by consulting a knowledge base, with well-calibrated artificial intelligence applications one could even solve the problem before escalating to an agent.

The next step is to include the descriptions given by users and use them to suggest solutions to the agents until the trust is high enough that the software independently suggests solutions to users or to route incidents.

In any case, intelligent routing reduces friction in the support process and, therefore, improves the experience of both the service agent and the customer. And most importantly: the result of these techniques in the positive experience is a measurable benefit to the business.