While it would be ideal if contact centers had all the time in the world to listen to recorded calls collected during the course of the quality monitoring process, it’s simply not possible. Contact centers still routinely practice call sampling, or monitoring one or two calls per month for each customer support representative to spot any egregious problems. The randomness of the sampling helps the quality process a little; after all, agents never know which calls are going to be monitored, so this may help them make an effort.
But what humans can’t do, machines can. Technologies that can sample more (think: all) calls can go a long way toward providing companies with a better sense of the state of their customer support. Speech analytics as a component of quality monitoring can “listen” to and report on a contact center’s calls for a variety of parameters: employee engagement, customer engagement, problematic calls, keyword spotting and more, according to David Ziv writing for Customer Experience Report. It’s meant to work in tandem with quality monitoring, either in software, premise-based solutions or third-party remote call monitoring.
“Speech analytics doesn’t necessarily replace the need for quality monitoring, call listening, evaluations or one-on-one coaching and training sessions,” he wrote. “With its ability to mine 100 percent of calls, it can, however, help make these programs much more efficient and effective.”
It improves quality. Sampling a few agent calls a month isn’t the best approach. Everyone has a bad day now and then, and listening to only a few calls of an agent who may take thousands a week isn’t statistically significant. Speech analytics can review far more calls and identify statistically significant knowledge, behavior or skill gaps across all calls.
It allows managers to find specific events. When a specific behavior is identified on a monitored call, managers can employ speech analytics to search through other calls from the same agent or others to determine if this is a repeat or a singular event, according to Ziv.
It makes it easy to search. Once upon a time, finding specific calls – ones that reference a certain product or service, for example, or a competitor – meant farming through them manually. Today’s speech analytics solutions allow users to search by keyword and find what they need quickly.
It removes the bias. While most contact centers calibrate quality monitoring processes regularly between evaluators to try and keep things fair, human bias will always creep in. According to Ziv, speech analytics technology is outstanding for removing this bias.
“Certain quality elements can be analyzed automatically with speech analytics and presented in a daily scorecard,” he wrote. “Some examples include: customer complements, complaints, requested escalations, confusion, requested callback, long holds and silences, mention of marketing promotions, proper call opening and closing, etc. By applying the exact same evaluation criteria to all calls and all agents, these elements of the quality scores become objective and statistically significant.”
Immediacy. When managers must evaluate calls manually, it’s often weeks or even months later when the agent receives feedback about certain calls. Speech analytics works at the speed of light (literally), and managers can receive information about problematic calls instantly. Many solutions allow managers to easily share those calls, sending them back to agents’ desktops with annotations added. (“Watch your script adherence on this greeting!”).
Whether you do your own quality monitoring or employ a third-party remote call recording services provider, speech analytics are relevant and can provide a better framework on which to hang a twenty-first century quality program for the contact center.