Blog
From Kickoff to Code: How We Built a World Cup App with GitHub Copilot
This idea, app, and post were created by the Innovation team of Randy Pagels, Trish Roberts, Zuhaer Zim, and Chad Cross from Xebia
At a Glance
- What we built: A World Cup Prediction App that lets Xebians predict match outcomes, compete on a global leaderboard, and connect with colleagues throughout the tournament.
- How we built it: The team used GitHub Copilot throughout the software development lifecycle, from application scaffolding and authentication to infrastructure, testing, and deployment.
- Technology stack: React, TypeScript, Vite, Tailwind CSS, .NET 8, Azure SQL, Entity Framework Core, Azure App Service, Azure Static Web Apps, Bicep, and GitHub Actions.
- How AI was used: GitHub Copilot acted as an AI development partner, accelerating implementation while engineers remained responsible for architecture, validation, testing, and technical decisions.
- The result: In a single Xebia Innovation Day, the team transformed an idea into a production-ready application for a global employee audience.
- Key takeaway: AI-assisted software development can help engineering teams move faster without replacing human expertise or compromising engineering standards.
How Xebia Turned a World Cup Idea Into an AI-Assisted Application
At Xebia, we set aside dedicated time for employees to step away from client work and focus on ideas that improve how we collaborate, innovate, and grow. We call this Innovation Day—an opportunity to explore new ideas, experiment with emerging technologies, and build meaningful solutions together.
During our first Innovation Day of 2026, our team set out to build something inspired by the upcoming FIFA World Cup. We wanted to create an experience that would bring our global workforce together through a shared passion for sports, connection, and a little friendly competition.
That idea became the World Cup Prediction App, a platform where Xebians around the world could predict match outcomes, compete on a global leaderboard, and engage with colleagues throughout the tournament. Built using AI Copilot Agents, the app combines the excitement of March Madness-style predictions with the scoring mechanics of Super Bowl squares, creating an interactive experience that's easy to join and fun to follow. Participants earn points for correctly predicting match winners, with additional points awarded for accurately predicting the final score.
What Does AI-Assisted Software Development Look Like in Practice?
While the app is a fun way to engage employees, its real value is demonstrating what modern AI-assisted software development looks like in practice. Built end-to-end with GitHub Copilot, it shows how AI can quickly turn a simple concept into a practical, scalable solution, spanning the frontend, backend services, and cloud infrastructure.
Before you picture someone typing "Build me a World Cup app" into a chat window and pressing Enter, let's be clear: that's not what happened. This wasn't a case of vibe coding or blindly accepting AI-generated code. Every feature, API endpoint, and infrastructure component began with deliberate, well-crafted prompts and was followed by careful validation, testing, and human review. GitHub Copilot wasn't a shortcut—it was a skilled development partner that accelerated the process while keeping engineers firmly in control.
How Did the Development Team Work With GitHub Copilot?
So what did that collaboration actually look like?
GitHub Copilot wasn't used to generate a single application from one prompt. Instead, it became a development partner, an AI teammate, throughout the entire project, helping accelerate implementation while our engineers guided the architecture, validated every decision, and refined the final solution.
What Technology Stack Powered the World Cup Prediction App?
Together, we built a modern cloud-native application using:
- Frontend: React, TypeScript, Vite, and Tailwind CSS
- Backend: .NET 8 Web API (C#)
- Database: Azure SQL with Entity Framework Core
- Hosting: Azure Static Web Apps for the frontend and Azure App Service for the API
- Infrastructure: Bicep Infrastructure as Code (IaC) for every Azure resource
- CI/CD: GitHub Actions to build, test, and deploy the application
How Does the World Cup Prediction App Architecture Work?
The Logic Flow
How Did GitHub Copilot Accelerate the Software Development Lifecycle?
Throughout the project, GitHub Copilot helped with far more than writing snippets of code. It assisted across the entire development lifecycle, including:
- Retrieving all 104 FIFA World Cup 2026 match schedules from the official tournament website, ensuring match data remained accurate and current.
- Scaffolding the complete Entity Framework Core data model, including users, matches, predictions, relationships, and database constraints.
- Generating the full JWT authentication flow, from user registration and login through securing the prediction APIs.
- Authoring every Bicep Infrastructure as Code file required to provision Azure App Service, Azure SQL, Key Vault, Application Insights, and supporting resources. Every Azure resource was deployed through code rather than manually configured.
- Building complete GitHub Actions workflows that automatically build, test, and deploy the application on every commit.
- Helping diagnose and resolve a production deployment issue encountered during development.
How Did Human Review Keep AI-Generated Code Reliable?
Of course, AI-generated code wasn't simply accepted at face value. Every feature, infrastructure definition, and deployment workflow was reviewed, tested, and validated by the engineering team.
GitHub Copilot also helped generate Playwright UI automation tests covering all scoring scenarios. The tests created fake matches, entered fake predictions and scores, verified the leaderboard logic, reset the application, and repeated the process for each scenario. This gave us confidence that the application would perform reliably before opening it to a global audience of highly competitive Xebians.
What Did We Learn From Building an Application With GitHub Copilot?
The biggest takeaway wasn't that GitHub Copilot replaced software engineers. It was that it eliminated much of the repetitive work between an idea and a production-ready implementation. The team remained responsible for the architecture, technical decisions, validation, and quality, while GitHub Copilot accelerated the execution.
The result? In a single Innovation Day, our team transformed an idea into a production-ready application that employees around the world could begin using immediately—a practical demonstration of how AI-assisted development can help teams deliver high-quality software faster without sacrificing engineering standards.
Written by
Randy Pagels
I am a DevOps Architect and Trainer at Xebia USA. I lead and educate customers in designing and implementing solutions that adhere to industry standards and DevOps best practices.
Our Ideas
Explore More Blogs
Contact