From support to root cause: what a CX person really does at Nubank

Written by: Mariana Foffa

Solving a customer’s problem is one thing. Understanding why that problem reached us is another.

The second part tends to be invisible from outside the area. When someone asks what a CX (Customer Experience) person does at Nubank, the most common answer is “they support customers”. And we do, end to end. But support is the beginning of the work, not the end of it.

The short answer is this: we resolve the case and then ask why it reached us. With that answer in hand, we look for ways to keep the problem from happening again.

In this text, I explain why that second question matters, how we work to answer it, how artificial intelligence is speeding up that work, and what this now demands from people in the area.

Why a case reaches us

If every process worked perfectly, customer support wouldn’t even need to exist. When someone contacts us, it’s because some journey didn’t work the way it should, or because that request isn’t yet available in the app for customers to resolve on their own.

And that’s okay. The future we’re after is a fully autonomous customer, who can do everything in the app and doesn’t need to ask for help. Until that future fully arrives, we’re here.

When the customer shows up, we resolve the case end to end: we follow it through to the finish and want them to leave delighted. But the work doesn’t end there. We also look at that case with a continuous-improvement mindset and ask ourselves what needs to change so the problem doesn’t happen again, or so that customer doesn’t need to contact us again.

I learned to look at things this way from my onboarding: seeing each customer as unique and, at the same time, looking at all cases as a whole. It’s in the whole that patterns appear, the things that repeat and are worth solving at the source.

Support as a source of intelligence

Many people see customer support as a cost to the company. I see it as an area that generates intelligence.

That difference shows in how the area is measured and what it chooses to prioritize. Our OKRs, which are the area’s goals, track customer satisfaction, team productivity and risk indicators, among others. Our projects, in turn, come from real problems that surface in support: a flow that needs to be structured, a recurring customer pain point, a process that doesn’t work well.

When a problem appears, there are two paths. If the solution depends on another area, such as product or engineering, we take the insight there, with data and in a structured way. If we can solve it ourselves, we do.

It’s on the second path that artificial intelligence is making the biggest difference.

Interim solutions until the permanent one arrives

Many improvements depend on development and take time to enter an engineering team’s roadmap (the list of priorities). That wait has a cost: until the permanent fix ships, the same problem keeps reaching support.

That’s where AI-powered interim solutions come in. CX teams have the autonomy to build temporary solutions that improve internal processes right away. They solve the problem now and show results, which also helps sustain the priority of the permanent solution when its turn comes.

An example: quality monitoring in second-level support

One example comes from my area, second-level support, which handles the most complex cases, those not resolved in the regular chat, email and phone channels.

An important part of our work there is quality monitoring: reviewing cases to check whether processes are being followed correctly. It was very manual and time-consuming, which limited how many cases we could review.

With the help of artificial intelligence, we developed in-house a solution using Apps Script (a Google tool for automating spreadsheets and documents) that sped up these reviews. Besides saving time, the solution began giving us tracking metrics, such as the time each reviewer spends on each evaluation.

Those metrics are the part that often goes unnoticed. By measuring the process, we identified the opportunity to expand sampling per agent and, with that, scale the volume of reviews.

And this reaches the end customer. When the process is followed correctly and the reviewed sample is larger, the chance of the customer leaving satisfied is higher. The internal improvement improves the experience of those who provide support and, in the end, of those who receive it.

The two skills the area now requires

Based on what I see today, CX professionals need to combine two skills.

The first is analytical vision: the ability to analyze data, identify patterns and understand the root cause of each problem. Without it, each case remains an isolated event, and the whole doesn’t become information.

The second is adopting and applying AI: the willingness to use artificial intelligence both to support analysis and to build solutions autonomously, without depending only on ready-made deliveries from other areas.

The two go together. Analytical vision defines which problem is worth solving, and applying AI shortens the time between identifying that problem and testing a solution.

CX as the starting point for improvements

Support is where many problems show up first. That’s why CX is also where many improvements begin.

A CX person at Nubank solves the customer’s problem, asks why it happened, and uses data and AI so it doesn’t happen again, whether by taking the insight to other areas or by solving it ourselves.

It’s work that starts with one case and ends with a process. And that’s why it’s worth asking, after every resolved case, why it needed to exist.You can do that today.