Generating AI-Powered CX Insights for the Frontline


Share on LinkedIn

In my role at Concentrix, I often talk to companies about their speech analytics efforts. What’s striking is how frequently these professionals express frustration and disappointment. Many feel oversold. Some are falling out of love with the technology. Most say they aren’t getting meaningful insights or quantified ROI despite having teams of analysts focused on it.

These same companies believed that a speech analytics platform would provide richer, deeper insights – but in reality it’s proven far more expensive than the original business case, with far fewer results than projected.

What’s causing these missed expectations? Is there a way to course-correct? Let’s explore.

The Root of the Problem

As brands fight for the attention of the always-on customer, organizations are making big bets on big data to understand needs, wants, preferences, and behaviors. Innovators in using big data to improve the customer experience are taking advantage of unstructured data, including the spoken word.

There are multiple benefits to using speech analytics in customer service, from performance management to cost containment. In the end, if it improves the customer experience it’s sure to improve the business.

So why do issues persist? A key part of the problem is that companies underestimate what’s needed to make technology effective. Early adopters of speech analytics as a supplement to quality monitoring quickly recognized that more data alone doesn’t improve interactions. Whether focused within a service center or broader behavioral journeys, the processes and skills placed around a speech implementation are critical to realizing business results.

Finding a Solution

The latest innovations in speech analytics emphasize the importance of having smart people actively managing inputs and outputs on an ongoing basis. The tools themselves are extremely robust but that means they’re not a plug-and-play option for IT to install and your team to noodle around in. To really achieve value, you have to know how to do it right and be willing to make the proper investments.

Too many business leaders generally focus on, “What can speech do for me?” instead of shifting focus to how speech, along with other analytics, can dig into specific problem areas. With the evolution of real-time squarely in our midst, the benefits of speech analytics are becoming clearer and more tangible.

Any software by itself will provide some insights. But the accuracy, relevancy, and actionability of underlying taxonomies takes both art and science. It requires professional staff experienced not only in speech technology in general, but also in broader operational and customer impacts. Only with a skilled team interfacing with the software on an ongoing basis can you hope to uncover patterns that you would otherwise have missed.

7 Ways to Activate Speech at the Frontline

Knowing how speech insight translates to daily frontline activities is key to success. Here we share several tips on how to get your speech program quickly producing a return on investment and informing frontline change.

1) Narrow your focus. It’s easy to get overwhelmed when starting off with speech. Avoid throwing a bunch of categorizations and every possible data element into analyses. The better approach is to start with a few key business questions, like why customers are defecting or why repeat contacts are increasing. If you’re unclear where to look, start with analysis of survey data to uncover drivers of customer issues. Develop an approach that drills into root cause and deploy actions from there.

2) Keep speech in context. Too often companies have a team of data scientists or AI specialists exclusively conducting speech analytics. Instead, analysts with experience in the business or the service center are better able to quickly get at relevancy and accuracy. They know how the environment works and are more adept at knowing what to look for. Today’s software doesn’t require a doctorate degree, so you shouldn’t either.

3) Look broad and deep. Speech is highly effective at quantifying higher-level insight but it won’t tell you what’s happening or why. By linking other data sources with speech data, analysts can develop a fuller picture and better isolate unexpected patterns. Machine learning and deep learning make these activities even easier.

4) Build trust in the data. Encourage operational teams to provide feedback on categories and flagged calls, particularly in the early stages. Involving them in speech from the start builds consensus and enables them to participate in calibration that ensures relevancy to the environment.

5) Move beyond dashboards. Speech technology providers want you to think dashboards are all that’s needed to improve performance. Reality says otherwise. When it comes to the frontline, role-based dashboards are highly effective at streamlining content and focusing on impact areas. But dashboards alone aren’t enough. No matter how great the dashboard, there’s still work to be done beyond graphical summaries.

6) Inspire action. Best practice requires building speech into daily action plans, coaching, team huddles, and recognition. These are proven hallmarks of companies successfully changing internal culture. Analysts can size business case for action; software can’t. Analysts can develop clear return on investment projections and attach financial impacts to a list of recommended action; software cannot. Without action there is no return on investment.

7) Don’t isolate speech. Remember that speech analysis is one of myriad inputs that, taken together, help inform frontline performance. Speech should be used to pinpoint problem areas as well as uncover positive performance. Creating an environment when speech is viewed positively will go a long way to inspiring staff.

In Summary

Not surprising to contact center leaders, one of the best sources of customer insights lies within the gigabytes of call recordings. In the past five years, there’s been an avalanche of companies trying to use speech technology for day-to-day agent compliance and performance. Placing speech automation on top of call recordings gives agents more feedback than ever about performance. To maximize the investment, blend speech into ongoing processes and frontline activities. Done right, it’ll yield deeper insights that your team can use in process improvements and targeted actioning.

Image Source: GettyImages-1016109338.jpg

Gina Massa
Gina Massa is Senior Director, Analytics at Concentrix. She has over 20 years’ experience helping companies use customer experience insights to drive improvements at the point of interaction. She leads a global team of analysts and CX experts focused on optimizing processes and performance.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here