Nubank: 60% de Cobertura QA, +10 Pontos tNPS | Birdie

About the Customer

Founded in 2013, Nubank has become the largest digital bank in the world, serving more than 100 million customers across Latin America.

Built on a mission to simplify financial lives and democratize access to banking, Nubank has set new standards for digital experience, customer trust, and operational excellence. From the start, Nubank’s success was rooted in its human approach to technology — or as they call it, the “Jeito Nu de Atender” (“The Nubank Way of Serving”). Every customer interaction is meant to feel human, empathetic, and empowering, even as the company scales across millions of conversations per month.

But with growth comes complexity — and maintaining quality at that scale required a new way of thinking.

Situation and Challenges

As Nubank’s customer base expanded, the operations and service teams faced a familiar yet critical challenge:
How do you ensure quality and consistency across millions of customer interactions, handled by thousands of agents, in multiple channels — and prove that quality is actually impacting business outcomes?

Their Quality Assurance (QA) process had long been a cornerstone of their service operation. Analysts would manually review interactions, evaluate them against predefined criteria, and share results with leadership. These reviews provided valuable qualitative insights — but they were slow, subjective, and difficult to connect with broader business metrics like Transactional NPS (tNPS) or churn.

The team also had a well-defined set of 16 evaluation criteria, including five representing their cultural pillars — the “Jeito Nu” principles.
However, they lacked visibility into one key question:

“Which specific customer service behaviors actually move the needle on satisfaction and loyalty?”
Without that clarity, QA could not influence business strategy. It remained a back-office function — valuable, but not yet strategic.

How Birdie Helped

That’s where Birdie came in.
Nubank partnered with Birdie to use its AI-driven Quality module, Agent QA — part of Birdie’s Customer Intelligence Platform that connects customer feedback, support interactions, and business outcomes.

The goal was clear:

• Scale quality evaluations from a small sample to the majority of interactions.
  
• Quantify the impact of each agent behavior on customer satisfaction.
  
• Transform QA into a proactive system that predicts and drives tNPS improvement.
  
  

AI at Scale: From Monitoring to Modeling

Using Birdie, Nubank was able to automatically evaluate tens of thousands of customer interactions across channels. Each interaction was analyzed against 16 quality behaviors, ranging from empathy and clarity to process accuracy and problem resolution. Birdie’s models then linked these behaviors to tNPS, uncovering not only correlations — but causal relationships between how agents behaved and how customers felt.

The findings surprised everyone. The team discovered that some of the “Jeito Nu” behaviors they believed most critical for customer satisfaction were not actually the biggest NPS drivers. Instead, a different combination of eight behaviors consistently predicted higher tNPS scores.

This insight marked a turning point.
“Birdie helped us move from measuring quality to managing impact,” said a Nubank Quality leader.
“For the first time, we could see exactly which actions from our agents made customers more loyal — and where to focus our coaching and operational energy.”

From Insights to Action

Birdie didn’t just deliver the analysis — it also operationalized it.
Insights were embedded into dashboards and alerts that helped QA leaders, team managers, and CX strategists act quickly.
They could now prioritize coaching on the behaviors that mattered most, track their improvement over time, and measure how each change influenced tNPS — all in near real time.
Within weeks, the team had turned static QA data into a continuous learning loop between agents, operations, and business metrics.

Results

The results redefined the role of QA at Nubank:

• Identified a direct causal relationship between quality behaviors and tNPS
• Discovered that only 8 of the 16 behaviors drove the majority of NPS impact
• Shifted the QA focus from 16 to 8 strategic behaviors that matter most
• Increased coverage from <5% to >60% of analyzed interactions
• Reduced time from evaluation to action from 2 weeks to <24 hours
• Expected tNPS uplift: +5 to +10 p.p.

This clarity transformed QA from an auditing process into a predictive engine for customer loyalty.

Beyond Metrics: A New Strategic Role for QA

These results didn’t just improve processes — they redefined the purpose of QA within Nubank.
Instead of acting as a control mechanism, QA became a strategic enabler of growth, helping the company make better decisions about training, resource allocation, and customer strategy.
With Birdie, Nubank’s QA function now serves as a predictive system for customer satisfaction — identifying friction before it affects customers and guiding operational excellence from the inside out.