Technology

Speech Analytics in 2019 and Beyond: With Change Comes Opportunity

Claire Bekken
Claire Bakken is a Senior Analyst at MainTrax

If you’ve been lucky enough to be in the audience when an engaging speaker hits the stage, you know that his or her first words come with authority. Because of that, everyone else’s ears perk up a little bit, the notebooks and pens come out to scribble down the firehose of information being shared, and each member of the audience feels a little more at peace knowing that they’re getting the answers that they came to the session to find. And, by the reaction of the rest of the group, they know they’re not alone.

That’s how our team at MainTrax felt when, earlier this year, Strategic Contact and Contact Center Pipeline released its annual survey that looked at the top challenges and priorities from nearly 300 participants in the contact center world. Just like being at a conference with a session that speaks right to you, our ears also perked up because the results reflected the trends we’ve been seeing over the past year. The theme forecasted this year was change, and as new strategies for tackling these challenges and priorities are being rolled out, one quote comes to mind: “If nothing changes, nothing changes.” Contact centers are screaming for change, and we’re here to say: Let speech analytics lead the way.

Why Speech and Text Analytics

Think about the last time one of your agents told you every single detail from all of his or her calls over the last month. Not just the metrics, but the why behind those metrics. Were you given honest reasons for unearthly long handle times… were agents shuffling around the knowledge management database trying to find answers? Or how about unbiased opinions on cancel rates… were they due to a product flaw or issues with customer service? Could better save approaches have been used? Or what may have caused the slightest change to an agent’s estimated average sentiment score?

Speech AnalyticsThese confessional conversations haven’t happened yet? If you don’t know the specifics behind the metrics and the conversations going on within your company, it’s hard to make a quality plan to move your goal forward in the most effective way possible.

The difference between hearing and listening in our personal relationships also exists in our relationship with data. To have successful personal relationships, you need to know who you’re communicating with, listen to what they say and take note of what their actions tell you. With speech and text analytics, you can do the same thing with your agents and customers. By looking at the actual data, by listening to the actual voice of the customer, and by actually knowing what an agent’s specific process is—down to the second, phrases used and emotion—these changes and priorities are easier to accomplish than you might think. All it takes is the simple shift from hearing to listening.

Let’s look at the top challenges and priorities found in the Strategic Contact report and how we’ve seen speech and text analytics guide companies facing similar issues to successful outcomes.

Top Challenge: Lack of and/or Bad Desktop Tools

The top contact center challenge listed for 2019 was desktop tools, which rose 9% over last year. As Strategic Contact Founder and President Lori Bocklund stated, “It undoubtedly ripples into other issues like handle times, training and even attrition.” So, how do you utilize one technology (speech and text analytics) to alleviate the challenges presented by another technology (desktop tools)? It all goes back to that one word: Listen.

Issues caused by the absence of the proper desktop tools can show up in many places, especially depending on the department and role. For agents, issues may surface while using the CRM database to pull up and verify customer info. Or, for the same team, issues may occur when the agent gets increasingly flustered while trying to solve a customer’s problem because they’re not familiar enough with a knowledge management platform to navigate it effectively. There may be trouble opening the new screen-sharing tool or difficulties even getting the customer to the right place to start the screen-sharing process. If the IVR misrouted a call, the agent may struggle to find the right department to send the customer or, adding even more handle time, they may use the knowledge management database to answer the customer’s question on a topic on which they have not been trained. The list goes on.

If your contact center has implemented certain desktop tools or computer telephony integration (CTI), maybe your agents or supervisors are receiving contradicting customer information when the “screen-popping” function occurs (in which ANI is used to identify the customer and instantly populate their data on the agent’s screen). This very technology was purchased to improve human productivity, yet is proving to distract some agents. These desktop tools were designed with good intentions but, like any useful technology, it’s vital that team members apply best practices for using and evaluating them.

This, again, is where the listening comes in. You may think: How do I listen to my desktop analytics? Fortunately, speech analytics will provide you with the critical insights that you need.

Using Speech Analytics to Identify Desktop Tool Issues and Solutions

As identified in Strategic Contact’s survey, the absence of properly functioning desktop tools often increases handle time. To give you an idea of how speech analytics could aid in this challenge, let’s look to a company that used its speech analytics software to identify the challenges that one of their desktop tools—screen sharing—was causing and how, by correctly identifying those challenges, they were able to turn them into opportunities.

The company had launched a strategy using screen sharing years ago, believing that implementing this specific desktop tool would encourage immediate collaboration and positively contribute to overall productivity and greater CSAT. The problem was, they never actually measured the overall experience of how the technology was truly affecting different variables (e.g., AHT, NTT, sentiment, etc.), and because of that, the desktop tool’s promising features turned into a challenge.

Because the company already had a speech analytics platform in place, queries and business rules were built to find the subset of calls for which this specific desktop tool was being used. Once those calls were identified, they discovered:

  • The exact time in the conversation the technology was first mentioned.
  • When the optimal time to start using the technology would have been.
  • Which agents were overusing the tool.
  • How long it took to launch the screen-sharing technology after the customer agreed.
  • How long it took for screen-sharing to be implemented.
  • Why the process may have been delayed
  • The percentage of proposed sessions that failed

Was this tool being used properly? What were the hang-ups during the process? The findings were enlightening (see Table 1).

  • Increased Sentiment Score. By listening to the actual phone calls, it was concluded that because the agents utilizing screen sharing answered customer’s questions to completion, a higher sentiment score was reflected (1.57 without screen sharing; 2.75 with screen sharing).
  • Reduced Repeat Calls. By comparing ANIs from one month to the next, the speech analytics tool revealed that customers who were assisted with a screen share didn’t call back nearly as often as callers who weren’t. In fact, the number of repeat calls was cut nearly in half—from almost 57% to 30%—which was estimated to save more than $41,000 a month.
  • Opportunities to Reduce Handle Time and non-talk time. At first glance, one might look at these statistics and conclude that screen sharing had an adverse effect on increased averge handle time (AHT) and non-talk time (NTT), but further analysis exposed opportunities to improve each metric: Almost four minutes could be shaved off each call if agents were to offer the screen-sharing option at a more opportune time, representing a potential savings of approximately $35,500 per month. Although there was an increase in the average handle time of screen-sharing calls, the growth in sentiment score and reduction in repeat calls justified the rise in call duration.
  • Reasons for Slowdowns. Furthermore, the data showed that 62% of screen-sharing calls had delays—a large reason the tool was a challenge. Based on an analysis of which factors caused a delay in the implementation of the screen share (see Figure 1), the company was able to gain an understanding of the challenges that agents and customers were experiencing while launching and using the tool; consequently, they are now considering different technical strategies within the tool that will streamline the process.

The company now has a concrete idea about why the screen-sharing tool is in fact more profitable, along with percentages of elements that are causing slowdowns during use of the technology. Having the desktop tool in place and accepting that it was causing unknown challenges wasn’t enough. By using speech analytics in this area, they were able to find specific ways to use the information from the analytics to actually substantiate screen sharing, in addition to discovering how to turn the challenge into opportunities like increased monthly savings, a rise in average customer sentiment, lower handle times while using the tool, and how to tweak the technology to avoid slowdowns.

Most challenges can be overcome by listening: Listening to your agents, but also quite literally listening to phone calls, which will reveal where the roadblocks are taking place to help you get a pulse on what can be done to turn the situation around. With ever-growing speech and text analytics capabilities, the old manual process of listening call-by-call (or worse—just assuming what’s on the calls) to detect this highly valuable information is history. You may just find that your biggest challenges quickly turn into your biggest opportunities.

Top Priority: Improve Employee Engagement and Empowerment

Employee engagement also took a leap toward the top of the Strategic Contact survey. And, like Bocklund stated in her analysis, “Some of the other highly rated items, including self-service, performance tools, reporting and analytics, knowledge management and process improvements may help achieve better engagement,” too. Companies are prioritizing a variety of things, with many roads leading back to employee engagement and empowerment. However, getting your employees and agents to accurately verbalize and identify where these changes need to start is tough.

Using Speech Analytics to Empower Agents

While speech analytics is widely known for the ability to “hear the voice of the customer,” it’s also effectively used to: (a) Hear the voice of the agent; and (b) capitalize on what was heard from the voice of the customer to empower agents to be fully engaged.

Take another company, a global surgical and consumer medical supplies provider, as an example. They had implemented a newly rolled-out initiative to guide agents toward better customer experiences. Along with reframing their performance (QA) measurements and coaching styles, the company wanted to use speech analytics to improve coaching efforts by pinpointing areas of high importance down to the exact phrases and tactics being used, highlighting the agents who were implementing the new methods successfully, and presenting reverse-engineer scenarios where agents were missing the mark so that supervisors could empower them to have stronger customer conversations.

Through their speech analytics tool, a customized application was built containing four master categories and 14 subcategories filled with more than 500 audited and highly valuable phrases. This allowed them to:

  • See which departments the coaching opportunities were occurring in and filter between the two via a “Department” master category.
  • Identify when the new initiative’s methods were being used successfully—and celebrate those agents by looking into subcategories such as “Positive Language” and “Advocacy/Empathy.”
  • Compensate for part of their current manual QA scorecard with subcategories like “Taking Ownership” and “Next Issue Resolution.”
  • Spot instances where negative behavior was occurring on phone calls, and address and coach these scenarios in a timely manner through subcategories like “Credit Card,” in which credit card numbers were being recorded against protocol, and “Lack of Knowledge,” which alerted supervisors becoming to training opportunities.
  • Combine different subcategories together to pinpoint a set of even more narrowed-down high-impact calls, such as:
    • When a customer called in confused and the agent used the new customer experience methods to successfully turn the situation around (“Customer Confusion” + “Positive Language” + “Advocacy/Empathy”)
    • Evidence of agent lack of knowledge combined with specific coaching opportunities from the new initiative (“Lack of Knowledge” + “Coaching Opportunities”)

Before long, one manager admitted to doing a “little happy dance” when she realized what speech analytics could actually do for this initiative. She understood that, by having this system in place, the company was able to measure the impact of the new coaching program on a much larger scale, drill into specific calls featuring opportunities for further training needs to equip their teams with knowledge and confidence, and understand which methods were being effectively adopted by agents so that they were empowered to do their job while giving customers the best possible experience.

Closing Thoughts

While the list of submitted challenges and priorities in the Strategic Contact report looked long, many overlapped. As such, by tackling one, others will inevitably be impacted. By using speech analytics to address one challenge or priority for the year, I think you may be pleasantly surprised to see your to-do list shrink as an effect. For example, addressing the issue of poor desktop tool performance will likely also alleviate some of the challenges with increasing handle time (No. 12 on the list), channel additions and/or integration (No. 6), and lack of and/or bad self-service (No. 9). Using speech and text analytics to improve employee engagement will inevitably lead to improvements in knowledge management access (No. 5 on the priorities list), organizational processes (No. 6), and recruiting and hiring practices (No. 15).

Speech analytics is a fantastic way to whittle down your priority list and get others in your organization excited about how it can help them check off their own priorities, thereby contributing to the greater good and ensuring greater collaboration within the company (No. 4 on challenges and No. 12 on priorities!).

If the theme this year was change, I can only imagine what the theme for 2020 will be for those contact centers that utilize the vast abilities that speech and text analytics offer to help your company achieve its goals. We may see a whole new list!


What’s Trending in Speech Analytics?

Like any product or service in the tech industry, things are always changing and developing in the speech analytics world, too. Here are some trending topics and movements that we’re seeing in the industry right now:

  • AI is continuously being used specifically with speech and text analytics software to improve insights and accuracy.
  • Voice interaction data is being pushed into data centers and consumed with BI tools.
  • Machine learning is now becoming a reality to improve speech analytics solutions.
  • SAaaS—Speech Analytics as a Service.
  • Speech and text analytics companies are acquiring smaller vertical-specific analytics companies.

Keep an eye out for these trends and start brainstorming ways in which you can begin utilizing these industry patterns with your own speech and text deployments!

– Reprinted with permission from Contact Center Pipeline, http://www.contactcenterpipeline.com

Related Articles

Back to top button