PryvX
Collaborative Fraud Prevention Platform
PryvX
is platform helps organisations to combat fraud and cybercrime through collaborative efforts by leveraging privacy enhancing technologies, such as federated learning.
Problem
In today's increasingly digital world, cybersecurity has become essential for individuals, businesses, and governments alike. As we rely more and more on technology and the internet, our personal information, financial data, and critical infrastructure are all at risk from cyberattacks.The issue is that each telco and bank is trying to tackle this independently.
Goal
Design a collaborative platform that helps organizations detect and prevent fraud together, while keeping sensitive data secure and private.
Research
Interviewed stakeholders to understand fraud detection challenges and pain points. Collected feedback from potential users to uncover needs around collaboration and privacy. Conducted benchmarking of existing fraud prevention tools to identify gaps and opportunities for a secure, cross-company solution.
The platform architecture map clearly illustrates that the primary end user action is to use APIs. There are several ways to get to the API. It is API itself, Shared database, Projects.
In addition, the user has the opportunity to collaborate with others, combine data to improve the quality of cyber security.
Process
Platform architecture
Design
Create a new project flow
Initially, the creation of a new project was presented as one large form. After competitor research and testing, it was found that dividing into steps would help users fill out the form faster by focusing on specific inputs.
Define project
The user interface allows users to name their projects, assign tasks, and select algorithms. To make the tasks more understandable, they are visually represented using illustrated cards.
Configure ML settings
The system comes with pre-configured settings as a starting point. Nevertheless, users have the flexibility to personalize these settings to tailor the algorithm's training process and achieve their desired outcomes.
Import dataset
The user has the flexibility to choose a specific date from the preloaded database to begin training and can customize the output by selecting the desired categories.
Collabaration
The platform allows users to invite other registered users to collaborate and contribute to improving the accuracy of machine learning outcomes.
Fraud detection flow
APIs serve as the primary function of the platform. Various APIs types are supported. To facilitate navigation, a basic word search and a filter are included (popular APIs are displayed as filter bubbles beneath the search bar). Users can choose to view APIs as a grid or list layout.
Reflection
Team
Outcomes
This project was developed in collaboration with a business analyst, frontend developer, backend developer, and tech lead. I was responsible for UX research, wireframing, UI design, and creating interactive prototypes.
Designing a collaborative fraud prevention platform required balancing security, privacy, and usability. Simplifying complex workflows and visualizing cross-organization data clearly turned a technically challenging system into an intuitive, trustworthy experience for users.
✱ Collaborative fraud detection platform enabled multiple organizations to share insights securely, improving detection efficiency.
✱ Privacy-preserving design reduced concerns around sensitive data, increasing stakeholder trust.
✱ Streamlined workflows and intuitive interface allowed users to analyze and act on fraud cases faster.
