AI-Enabled Web App Reduces Financial Fraud

Scalable case management solution customizable to any industry

AI-Enabled Web App Reduces Financial Fraud

Story Snapshot

A financial services AI company needed to replace an existing case management product with a modern design, thoughtful user experience, and adaptability to many industries.

Results At-A-Glance

  • Reduced time to complete core tasks
  • Reduced complexity of customization
  • Implemented a modern UI with real time configuration
  • Created automatic workflows initiated by user’s actions
  • Rethought existing data architecture for scalability
  • Created an architecture that can be deployed to any cloud provider or on premises

Key Services

  • Heuristic evaluation
  • Usability testing, user proxy workshops
  • Persona and journey map creation
  • Web design and development

The Challenge: Buried Functionality, Lost Sales

A financial artificial intelligence company needed a complete refresh of their flagship transaction fraud case management system. The current app, which helped detect financial fraud for the company’s end customers, had an outdated user interface. Key features buried deep in the application. These obstacles led to lost sales opportunities and frustrated customers who weren’t seeing the value in the product.

The client brought in our team to help do a top-down overhaul starting with a human-centered approach.

Our Process: Human-Centered Design Agile

Our team started its discovery process interviewing client employees and key customers to get an understanding of their workflows and frustrations with the current product. Through this, we learned core tasks of managing cases were difficult for analysts and customizing the application for different customers was too complex.

We then worked with the stakeholders in story mapping sessions to curate a list of core features the application needed to have to service the different user personas. With the story map in place, we built an initial end-to-end case management solution that showed the entire workflow and laid the microservice groundwork for the client’s development teams to join us and help iterate and refine different aspects of the product.

We worked with the client’s customer success team to ensure the application allowed them to easily customize the application for specific industries. The design also emphasized ease of use for end users and analysts - so fraud can be identified faster. This focus on the end user experience allowed us to bubble up core functionality that was previously buried behind multiple menus and a 26-field search page. 

Now, the user logs in and is automatically shown the cases assigned to them, so they can start work immediately.

The case screen went from 13 tabs to three with a hierarchy of content. When the user arrives on the case screen, they’re taken to the investigate page, so they can jump right into getting context around why the case is potentially fraudulent.

Through our HCDAgile methodology, we worked with other designers to create a design system that applied a consistent look and feel across the entire application and set new brand and interaction standards for all of the client’s applications going forward. The design team worked to stay two sprints ahead of the development team to flesh out the experience in detail just in time for the developers to build it.

"This is so much better than what we have now. It makes so much more sense and is so much easier to use." 

- Customer Solutions Engineer, Financial Services AI Company

The Solution: A Scalable, Maintainable New Web App

We built a modern, scalable, customizable case management web application that focuses on eliminating end user friction to detect fraud faster and enable customization and automation. It can be deployed to any cloud or on-premise infrastructure. Given this infrastructure variability, we used Terraform to set the foundation of declarative scripts that could deploy the entire application stack in an idempotent manner.

Postgress was kept as the database since there were many years of experience with it at the client and the benefits of switching to another database didn’t outweigh the costs. Node was chosen as the ORM and API layer. This wasn’t a CPU-intensive product and mainly focused on writing and reading from a database so it was chosen over the existing Java/SpringBoot middleware stack.

One of the goals of this project was to lay the foundation for cases to be updated automatically through machine learning and, later, artificial intelligence. This meant a queuing system was needed to receive messages with information on what changes needed to be made to what cases. RabbitMQ was chosen because of its ability to scale subscribers quickly and its great library support.

For the front end, the React framework was used because the client already had a lot of infrastructure and knowledge around the framework through other products created at the company. There was also a component library based on an older design system that could be visually overhauled and leveraged throughout the application.

The Outcomes: An Easier Experience, Scalable Functionality

Modernized Design

With the client’s development teams, we created a product which sets the bar for a superior user experience going forward. By leveraging an architecture that was ready to be integrated with the client’s machine learning and artificial intelligence engines, the solution is easy for the client’s teams to maintain and scale long term. 

Thoughtful User Experience

The app was redesigned to detect fraud faster. For the client’s customers, we focused on bringing the core functionality to the forefront of the application, reducing the time it took users to get their job done. By streamlining case investigation, the app refocuses on the most valuable and urgent functionality, making it easier to use but also sell to new customers. 

Easily Customizable

The application is customizable from the bottom-up with its cloud-agnostic approach and scalable architecture. It allows for customizable automated workflows and UI where fields can be added or removed and display format of each field can be personalized to each industry or customer. This also made it easier to land new customers and for the financial services AI company to grow. 


Ready to write your own success story?

1904labs has the expertise and experience to tackle your toughest problems.

Request a strategy session