AI-powered sentiment in action: Creating happier agents, happier customers at PayPal and T-Mobile

Customer sentiment can change quickly and unexpectedly, veering from delight to disappointment as an interaction or brand experience fails to deliver as expected.

When negative sentiment occurs, the next sound you often hear is customers clicking over to a competitor—and your customer satisfaction (CSAT) and customer loyalty dropping.

On the other hand, positive sentiment can have a very positive impact on your business. This is why understanding and taking action on customer sentiment is vital.

What kind of quantitative benefits can you expect?

NiCE analyzed of billions of conversations across thousands of organizations—ranging from startups to industry giants—and found that those brands that excel in customer sentiment outperform their peers by an impressive 43 percentage points in stock returns over five years.

In this article, we’ll talk about why more companies are adopting AI-enabled sentiment solutions and share how two top global brands, T-Mobile and PayPal, are using AI-powered sentiment to improve agent training and CSAT.

More companies are adopting AI sentiment solutions

Across industries, AI-powered customer sentiment is helping top brands accurately measure and act on customer sentiment.

With a better understanding of customer sentiment—and how well agents are exhibiting the key behaviors that impact sentiment—CX team leaders can provide targeted coaching that helps agents deliver positive, meaningful experiences that customers appreciate and value—the kind of CX that drives long-term customer loyalty, and increased revenue.

That’s why companies that are looking to maintain or gain a competitive edge are increasingly adopting sentiment solutions.

In a 2024 global study from Metrigy, 32.6% of respondents said that they were currently using sentiment—a 17% increase since 2023. Another 59.9% said that they were either evaluating or planning to use it in 2024-2025+^[1].

NiCE sentiment is derived with a machine learning (AI) model trained with a large, comprehensive CX dataset that measures whether a customer interaction is positive, negative, or neutral on a relational scale. Sentiment provides a primary pillar for agent coaching, empowering managers to:

• See where each team member ranks at a glance
• Make comparisons among individuals in a team, across teams, or even across sites
• Easily identify improvement opportunities and provide personalized, targeted coaching
• Measure improvement over time

Two real-world examples of AI-powered sentiment in action

In a recent webinar, BreShawna Briggs, Manager of Contact Interaction Analytics at T-Mobile, and Jared Shumway, Manager of Interaction Analytics Development at PayPal, shared how their teams are analyzing and taking action on sentiment to improve agent coaching and well-being—to drive higher sentiment scores, and related positive business impacts.

Sentiment as a leading CSAT indicator at T-Mobile

We asked BreShawna and Jared what prompted them to get started with sentiment data at their organizations and what their main goals and objectives are now.

BreShawna noted that T-Mobile has been using sentiment since 2017 and it’s now an expected, integral part of the any major company initiative, because it provides a leading CSAT indicator, in conjunction with valuable direct customer feedback.

For example, after a contact center interaction, PayPal texts customers a transactional survey asking whether their problem was resolved. BreShawna said that sentiment gives them a nearly real-time gauge of how an interaction went—even while they’re waiting to receive a survey response, or if they never receive one, so they can take appropriate action.

To measure voice of the customer (VOC) more broadly, T-Mobile also randomly surveys customers, as BreShawna says, “just as a pulse check to gauge how are things going, whether or not they’ve contacted the company in the recent past.”

Of course, T-Mobile would like to see positive sentiment on every interaction, but BreShawna knows that’s not realistic, so their top objective is improvement, noting: “For us, sentiment is about: Are we getting better? We always want to be shifting if we start negative.”

T-Mobile company leaders take sentiment seriously and have dashboards with daily and weekly aggregate views of volume and where sentiment is tracking. “It's always about: Are we taking care of the customer? That is our North Star,” she said.

Watch the on-demand webinar to learn how T-Mobile used sentiment in a recent iPhone launch.

Identifying training opportunities and customer sticking points at PayPal

PayPal is primarily using sentiment to uncover coaching opportunities, so they look at sentiment in the aggregate and how an agent’s scores fit within the distribution. They’re also working on ways to operationalize sentiment data, and have had early success in their VOC program gauging customer reactions to product launches.

Jared describes sentiment as a “dowsing rod” that helps PayPal supervisors know where they should start digging. He notes that he’s “seen some eye-catching numbers” where 70% of an agent's calls are extremely negative and others whose calls are 80% very positive, which helps them identify best practices. Leaders drill down into the calls that led to really negative sentiment, listen to the recordings, and provide targeted coaching to help improve agent performance.

Sentiment analysis also helps PayPal pinpoint company policies that aren’t sitting well with customers. Jared notes: “People have been really impressed with understanding that this sentiment is telling us: ‘Here are some sticking points’, so let's change the approach or policies behind this so that we can help turn these people that are in the neutral or are negative into positive.”

Watch the on-demand webinar to learn more about PayPal’s data-driven approach to sentiment and coaching methodologies.

Sentiment’s impact on CX and bottom-line metrics

From better agent performance to lower costs, AI-powered sentiment is delivering a cascade of CX and business benefits at PayPal and T-Mobile.

Jared notes that at PayPal: “We’ve shown that when you focus on sentiment, you get better results…we’ve had some of the most incredible results and it was all measurable as an output from sentiment.”

BreShawna notes that at T-Mobile, because AI-powered sentiment evaluates all interactions, rather than just a 30- or even 100-call sample, “it has enabled us to significantly reduce manual call listening activities and lower operating costs.”

These results are not surprising. NiCE research^[2] confirms the positive impact of using sentiment to coach agents to better performance.

We compared the top-performing 10% of agents and bottom 10% of agents in our global study and found that better behavioral performance results in stronger bottom-line metrics, including:

• 256% More positive sentiment
• 25% Shorter calls
• 53% Less non-talk time
• 53% Fewer repeat contacts

Key behaviors that impact customer sentiment

Based on years of research, testing, and analysis, NiCE has identified nine key behaviors that have been proven to impact customer sentiment:

1. Taking appropriate action
2. Demonstrating ownership
3. Effective questioning
4. Promoting self-service
5. Active listening
6. Being empathetic
7. Acknowledging loyalty
8. Setting expectations
9. Building rapport

When agents demonstrate these behaviors effectively, sentiment becomes more positive—and when they don’t, it goes negative.

NiCE Enlighten AI for CSAT includes behavioral models that objectively measure the nine key behaviors, opening the door to a new dimension of understanding which behaviors most impact sentiment scores, and helping both supervisors and agents focus on the behaviors that matter most.

Creating a personal connection with customers, responding to customer needs in real time, and delivering an exceptional experience all have a positive impact on customer sentiment.

In fact, NiCE research and analysis found that on a 0-100 scale, customer sentiment soars by 50 points as agents demonstrate more positive behaviors. This emphasizes the impact of soft skills, and reinforces why good agent training and on-going coaching are vital.

Don’t be left behind

As the research shows, and real-world experiences at T-Mobile and PayPal confirm, companies that are using AI-powered sentiment analysis, and reinforce key soft skills through effective coaching, see their efforts pay off with both CX KPI and bottom-line gains.

So, if you’re not already looking into AI-enabled sentiment, your organization will quickly be left behind. Reach out to a NiCE specialist today to start a conversation about how to use sentiment more effectively in your organization.

Watch the full webinar on-demand

_{[1] Metrigy: AI for Business Success Global Study (2024)
[2] NiCE: The State of CX: Real Insights from the Largest CX Dataset (2024)}