In the world of customer service, quality and experience are not only the goals – they’re the currency by which organizations show value to their customers. And while that all sounds great… it’s not easy to measure. Traditional quality assurance (QA) processes involved spot checking call, chat and email transcripts – well, actually a lot of that still goes on today. But the main idea was that a random sampling would be good enough to measure quality. Automated QA promises to make that a thing of the past. But does it actually work? And if so, how?
Let’s start with the basics of Automated QA in customer service
Automated Quality Assurance (QA) in customer service refers to the use of automation tools and software to assess, monitor, and evaluate the quality of interactions between customer service representatives and customers. The idea behind this approach is to streamline the evaluation process, reduce bias within the assessment process, and improve the overall quality of customer service interactions.
As I said above, automated QA includes all elements of customer experience, from phone conversations to live chats to email correspondences, and more. The actual assessment is performed by a QA analyst using predefined criteria and key performance indicators (KPIs) to assess the effectiveness of customer service interactions. This is typically done using a scorecard, which is really just a list of various criteria covering factors like response time, problem resolution, adherence to scripts, and overall customer satisfaction.
Benefits of Automated QA
If you don’t already have a QA program in place, you may be thinking “do I really need I really need one – let alone an ‘Automated QA’ program?’” The short answer is… that depends. There are a number of direct and indirect benefits to QA, which I’ve laid out below:
- Consistency: Automated QA ensures that customer service representatives adhere to predefined standards and consistently provide quality service, reducing the risk of human variability. This can identify a problem you don’t know you have and otherwise would take a while to find.
- Efficiency: Automation tools enable rapid and comprehensive evaluation of customer interactions, significantly reducing the time required for quality assessments. As above, it shortens the time to issue identification and can improve agent ramp time as well.
- Objective Assessment: Automated QA relies on predefined criteria, making the evaluation process objective and free from human bias. In addition, automated QA tools that have a universal scoring criteria, like a CQ score, analyze each conversation equally before it’s even seen by an analyst. (Never heard of a CQ score?! Read more here.)
- Real-Time Feedback: With the ability to assess interactions in real-time, Automated QA allows for immediate feedback and the opportunity for timely coaching and improvement. For issues like profanity or other escalation this can be critical, especially if you have remote agents you don’t monitor directly.
- Scalability: Automated QA can efficiently scale to handle a large volume of customer interactions, making it ideal for businesses with high customer service demands. Similar to agent efficiency, it enables your QA analysts to do more evaluations, at higher quality, in less time.
- Data-Driven Insights: This is the latest trend in QA and perhaps the most beneficial to businesses as a whole. The data collected through Automated QA provides valuable insights for improving customer service operations, identifying training needs, and enhancing customer experiences, but also more. This can also identify product issues, policies that your customers don’t understand, or other points of friction in the customer journey. These are the elements you may never find in a survey, because customers may not know how to articulate them until they have a problem.
But the main reason the importance of QA depends on each individual organization’s perspective on customer experience (CX). Do you need to keep customers coming back, so consistency is important? Are you growing, so efficiency and scalability are critical? Even if CX isn’t how you keep your customers with you, some of the efficiency and feedback may have more short-term cost benefits. But which combination of the above is most important is unique to your business.
The different tools, methods, and approaches to Automated QA
Here’s where it gets interesting. While all the above sounds relatively straightforward, how companies within the “automated QA” category define what they do is wildly different. There are really four main approaches to QA:
- Manual QA: If you’re reading through transcripts and using a Google Form, Google Sheet or Google Doc (or their Microsoft equivalent), you might be doing manual QA. This approach is as the name implies, manual; taking a lot of time and only really sampling between 2-5% of all customer interactions. Your analysts use some kind of doc, form or spreadsheet to score each conversation and that data… goes somewhere never to be seen again. The main downside is, you don’t have an accurate view of both agents and your entire customer service operation. And for phone calls, your QA analysts may be spending a lot of time listening to capture what they need. We’ve heard of businesses spending 20 minutes to score a six minute phone call. Yikes.
- “Digital QA”: While this sounds like a step up, it’s really just a more advanced method of manual QA. There is still sampling, but customer interactions like phone calls, chats, and emails are assigned automatically instead of picked by a QA Manager. This is what constitutes the “automated” label. Other than that, these tools provide minimal reporting, perhaps some trend data based on keyword searching, but nothing in the realm of what we would call “AI” today.
- “AI QA”: Ah, AI. All your QA problems solved with no worries! What could go wrong… Yes, it does sound too good to be true doesn’t it. These tools may sound new, but most have been around for a few years, meaning they’ve built their platforms on what is now a dated approach. They also take a very long time to “learn” what good QA looks like, including multi-month implementation and a process of updating each individual criteria or line item within a scorecard. Most incorporate human feedback in order to train the model, so that is partially why the AI training is so tedious. And organizations often remove human analysts once the AI is trained, so it’s incredibly hard to understand whether everything is being scored correctly. One other note on these tools is their focus on either specific channels – often voice – and use cases – many are more for sales than customer service.
- “AI-Enhanced QA”: If you thought AI QA was the pinnacle, I have one more for you. This approach balances what AI is good at, namely seeing trends within large sets of data, and what QA analysts are good at, understanding context and softer skills. The methodology here is to analyze all interactions using AI, and provide recommendations to analysts to speed reviews. It’s really the best of both worlds: broad insights and objective analysis from AI and the understanding of human customer service experts.
Ready to see if Automated QA is for your organization?
If you see the benefits of continuous improvement, more consistent agent performance, and an easy way to get insight across your customer service operations, automated QA is a great way to get there. To learn more about how Loris approaches Automated QA, head over to our product page.
Image by Gerd Altmann.