One of the core maxims of Product Design is that every product should be easy to use. The word "intuitive" is constantly cited as the gold standard for good design. But what happens when a product is so easy to use that it ends up creating new problems? This case study explores that very question.

To understand how we reached that point, a little context is needed. At Brick, our initial goal was to develop an anti-fraud platform to solve a pain point in Brazil's car rental market. The user screening process was notoriously slow, as employees had to manually check multiple data sources; it was subjective, as there were no clear rules defining a "good renter", leaving the decision to each analyst's personal judgment; and it was prone to human error, as the high volume of analyses had to be done under pressure, often with the customer waiting at the counter.

Everyone knows a traffic light

Based on these problems, we developed the platform with three pillars in mind: relevant information, excellent data sources, and an easy-to-use system. After speaking with numerous clients to map out the crucial data and with the data team securing solid sources, we focused on the interface. Centralizing everything would already be a huge productivity gain, but we wanted to go a step further. The key insight was to use the logic of a traffic light: if the information was positive, the summary would turn green; if negative, red; if it required attention, yellow; and if it was neutral, black. It was a major success. Analyses became much faster, our clients' operations improved, and our user base began to grow.

Blinded by the Green Light

After a while, we started getting some interesting feedback. The system was so easy to use that it generated an unexpected behavior: rental agency employees were no longer analyzing user profiles in-depth. They would simply see if everything was green and approve the rental in seconds, causing them to overlook important information that could be critical. This led to an increase in fraud—a paradoxical effect of the efficiency we had created.

We could have blamed this on the rental companies' training processes, but that would mean shifting the blame, not solving the problem for anyone. So, we decided to act on a three-pronged solution.

Our short-term action was to review all alerts and raise the bar for what we considered a good customer, a task done in partnership with the rental companies. For the mid-term, the plan was to automate the decision-making process for the most obvious cases (good or bad), ensuring that only the profiles that truly required a critical eye were escalated for human review. Finally, our long-term vision was to re-architect the entire analysis logic, allowing each risk manager to create their own custom approval algorithm with even more features.

Where we're going, we don't need traffic lights

With this strategy, we mitigated the immediate risks and ensured operational stability for our clients while the team developed definitive solutions. The automated analysis is already a consolidated success, but it's the evolution of the custom algorithm tool that represents our greatest leap forward, empowering rental companies to build a more secure future and redefining what's possible in fraud prevention in Brazil.