When redesigning navigation wasn't the real problem
The BTG Pactual Bank customer support portal brought together accounts, cards, credit, and onboarding in one place. Our initial hypothesis was about visual organization. User testing proved us wrong.
BTG Pactual Bank
Product Design Lead
2023
~35%
in escalation reduction
+2.1k
users impacted
+4
journeys delivered
01
Context
PLATFORM
One of BTG Pactual Bank's main customer support platforms
SYMPTOM
Scattered information and limited autonomy to resolve cases
GOAL
Reduce operational dependencies and make support workflows faster
BEFORE
Onboarding, transactions, cards, and loans — all in one place, all mixed together.
02
Evidence of opportunities for Level 1 resolution
I started by comparing the most frequent occurrences handled by Level 1 and Level 2 support. The differences between the rankings helped us question whether every escalation was actually necessary.
Together with the support team, we reviewed which cases genuinely required Level 2 expertise and which could be resolved naturally at Level 1 with the right information, permissions, and tools. This revealed opportunities to reduce unnecessary handoffs and improve first-contact resolution.
MOST FREQUENT ISSUES HANDLED BY LEVEL 1
Personal Credit
1.00x
Biometric
0.93x
Credit Limit Request
0.79x
New Card Request
0.67x
Limit Balance
0.64x
MOST FREQUENT ISSUES HANDLED BY LEVEL 2
Biometric
1.00x
New Card Request
0.94x
Credit Reassessment
0.77x
Onboarding
0.66x
Refinancig
0.55x
03
Testing revealed what was really driving escalations
Quantitative data showed where escalations were concentrated, but not why they were happening.
I then used interviews and usability testing to investigate the causes behind them.
DECISION TURNING POINT
I initially focused on helping agents find information.
Testing showed that the greater opportunity was enabling them to act on it.
INITIAL HYPOTHESIS
Agents were escalating cases because information was scattered, difficult to find, and lacked a clear visual hierarchy. I believed reorganizing the experience would reduce these dependencies.
WHAT TESTING REVEALED
Agents could usually find the information they needed. The main barrier was that they lacked critical data, permissions, and actions to complete the resolution independently.
This shifted the project from an information architecture redesign toward improving operational autonomy. Three findings shaped the new direction:
01 - Critical information had low visibility
High impact
02 - Essential data, permissions, and actions were missing
High impact
Medium impact
04
Hard calls
With limited time, we chose depth over coverage.
WHAT WE LEFT BEHIND
Reduced scope — we prioritized only the journeys with the highest operational dependency
WHAT WE GAINED
Focus on the areas driving the most escalations: onboarding, cards, and financing
I reallocated three designers already working on the product to reduce onboarding time and speed up delivery.
05
Building a shared foundation
The project exposed another gap: there were no established standards for internal web products.
Alongside the deliverables, I defined a shared foundation of navigation patterns, layouts, and components so different teams could work with consistency.
This became a prerequisite for scaling delivery, reducing development rework, and making QA validation more predictable.
06
Solution
BEFORE
AFTER
