Last March, the pandemic erupted and put contact centers to the test. They transitioned almost instantly to working from home (WFH) and confronted surges of customer service requests. While most stepped up to the challenge of handling the masses, one year later, many are still wrestling with volumes of inquiries growing faster than they can cope with. I now think that customer service faces a systemic capacity issue. In this two-part series, I’ll explain what this challenge means and walk you through technology and organizational levers to address it.
Capacity issue?
In the early days of the WFH transition, the main obstacle was setting up a remote work environment for agents. Most businesses were able to do it swiftly, but not all. While the pandemic has reduced churn, it has also made hiring, onboarding, and training more challenging. Net-net, things have been a wash, leaving customer service with roughly the same agent capacity. At the same time, the sheer increase in customer inquiries hasn’t stopped. In the beginning, the issue was about customers checking on COVID-related issues. Now, the world has become more digitally-inclined but even digital processes still require human assistance.
Digital to the rescue
Businesses have turned to self-service and automation to reduce the volume of human interactions. The first step to call avoidance involves offering customers the ability to help themselves on the web or mobile. Today’s state-of-the-art web self-service includes searchable help centers and on-page assistance. Help centers provide a one-stop location for all questions customers might have without having to navigate a site or a mobile application. Help centers have also been evolving from frequently asked questions (FAQs) to searchable knowledge bases of articles and can be augmented with forums where customers help each other.
Although web and mobile apps keep on improving their usability and offering more information, they aren’t enough. Enterprises I speak with consistently share that 50-70% of the inquiries they receive are coming from frustrated customers unable to resolve their issues online. The main problem with these calls is that their context gets lost because they’re anonymous and need to be resumed from scratch.
Businesses have deployed chat to provide contextual help but staffing remains a challenge. Teams are small, which creates long wait times and makes the sourcing of all the required skill sets difficult. Bots and conversational assistants can front-end chat sessions, provide some answers, and alleviate poor response times. While an excellent step, the model works as long as you can provide easy access to a human. Otherwise, the effort put in by customers is lost at the expense of their experience.
Persistent chat and asynchronous messaging, which allow the context of a conversation to be carried forward promise to improve the chat experience. That’s because customers can ask a question and “come back later” for their answer. Successes such as T-Mobile messaging service show that customers are fine waiting a little longer in exchange for getting their time back. But businesses still struggle to set and meet expectations on wait times and provide visibility into pending inquiries which leads to customers asking the same question on other channels.
Callback solutions offer another path forward and have evolved beyond giving customers the option to be called as soon as an agent is available instead of staying on hold. Callbacks can be offered online to capture the full context of the request. They can be combined with scheduling and/or notifications to offer great experiences. They can also be grouped and prioritized to be handled more efficiently.
Voice self-service
Voice self-service also needs to be optimized. Interactive voice responses (IVRs) already process about 50% of incoming calls and route a large chunk of the remaining ones. The level of automation IVRs can achieve is rather good. It’s the time and effort they demand from customers and the forced navigation through rigid menus that have hindered their ability to reduce calls to agents.
Speech is now a mature technology that not only makes interaction a better experience but also flattens menu hierarchies, giving faster access to options. Authentication is another mature technology that reduces the time spent in IVRs and avoids customers' painful repeats. Eventually, conversational AI can be used in the voice channel. The technology has made leaps and bounds and can push the envelope of what can be done in an IVR. Voice assistants can be deployed iteratively by focusing on specific branches of voice applications or segments of callers. More natural dialogs create better opportunities for deflection than today’s voice messages that tell callers to visit a website. With people mostly using mobile phones, SMS is an effective deflection channel to consider. Callers in queues can be offered to receive a text SMS. Messages guide customers to better destinations for getting their questions addressed and offer alternate options that put callers in control.
Speech can be combined with natural language processing to uncover intents and tag calls, which helps route them more precisely. Tagged interactions can better be routed to the best-skilled agents who can handle them faster. The instinct of many practitioners experiencing heavy call volumes is to create large pools of agents. Unfortunately, the larger the group, the broader the variability of handling times. It’s preferable to target smaller groups of experts who can handle interactions in a fraction of the time required by less proficient ones.
Contact centers faced with interaction volumes growing faster than their capacity have been developing self-help options for their customers. The issue is shifting to better handle inquiries that cannot be fully resolved using self-service. The next frontier is about assisting as an extension of self-service, with a seamless transition to agents. In part two, I will discuss how to optimize your talent pool and channel mix.
The article was initially published on No Jitter