Speech Analytics is about the business’s goal and what the business is set up to achieve. No matter what, typically, there’s going to be some human element. But it really just depends on what you’re trying to achieve. The reason why we really love speech analytics is that it listens to all calls. Before when it was a manual process, we were listening to probably 1% or less than 1% of calls. Now with speech analytics, we are listening to all of our calls based on categories of like, compliance, efficiency, things of that nature, and so on.
So on a score and category basis, we’re using those to replace QA and just say, “This person, based on our scoring, has a 98–99% compliance score,” and that takes away the human element of it. However, there might be somebody who knows they are hitting compliance points every time. Management or an analyst is going to have to look into that and say, “let’s go and look into their calls and see if they’re actually doing it.”
“At RTR, for example, we have a Do Not Call Scrub, which is amazing, and that’s very important to us. If someone says, ‘cease calling me, I don’t want to hear from you anymore, do not call this number anymore, or this is simply a wrong number, I am not this person,’ we need to make sure that we process that request. So what we’ll do is every day, we’ll have a report from CallMiner, and we’ll make sure that the system is updated, and that the agent handled that call right, or that they handled it correctly. So what that means is, we’ll go in and check, did they check the box Do not call?” – Andrea Jaegerman, RTR
Human Interaction & Speech Analytics Blend
We need human interaction, but you would never be able to have those results if you didn’t have Speech Analytics. Otherwise, you’d just be listening to random calls you selected, and if one happened to be a Do not call type of call, then you go and check it out. But using speech analytics, you can get all calls and listen to everything. So it’s basically a blend.
Basically, speech analytics automates the mundane. In the past contact centers had to listen to calls manually before speech analytics was out there. Speech analytics empowers Quality Assurance (QA), and the human element will continue to remain within. But what automating the speech analytics and doing all of that automatic scoring does is, now when you’re having that performance coaching discussion with that Agent, you are doing targeted listening. Instead of having to listen to a bunch of manual calls, you can now flag the superlatives.
Recognize Employee Successes
Speech Analytics isn’t just about the things the agents are doing wrong, you can take speech analytics and use it to compliment and recognize when employees are doing a great job and hitting all of those critical KPIs. It empowers and enables a better form of human connection between the coach and the agent. Whether it’s a quality assurance person or performance coach, speech analytics can provide positive reinforcement. You can say, “Hey, I see that you’re doing really great on call openings, you’re polite, and you’re empathetic. But your call closing or summary scores are a little bit low, so let’s focus on what we could do to improve that.”
Focus on Specific Areas
Speech analytics allows you to focus and really targeting specific areas. If somebody’s doing a great job with politeness, why are we talking about politeness? You can compliment them on it, but go on to the things that are really important to focus on. So it really empowers targeted coaching.
Automate the Mundane
Speech analytics is not going to replace human QA, but it does shift the focus of what QA does. It’s like a sort of modern manufacturing, where robotics does a lot of work, but you still need somebody to man the machines and make sure it’s working properly. The future of speech analytics is to automate the mundane, but you still need the QA to analyze that data, provide the coaching, dig in the things, and make sure everything’s working right.
With speech analytics, you can take very complicated scorecards that contain a lot of items that can’t be automated, and speech-enhanced QA. Instead of automating the scoring, put in tokens where certain events happen. Then a QA specialist can click through those tokens and jump to those parts of the call and check that if a person gave an objection. You are able to jump right to that section and listen for the rebuttal to see if they got a proper rebuttal, and they can score appropriately there. QA folks love this system. In fact, they’ve said their quality of life improved, and they’ve said that they can’t imagine doing QA the old-fashioned way. Because it becomes more of an investigator and less of a grater, and it’s just a little bit more engaging for them to do that.