Case Study: Bag Index

Client: Allan Bell *FDE: Arkash Jain

1. Introduction

Overview:

Bag Index is a web platform designed to help college athletes understand their financial worth in the rapidly evolving NIL (Name, Image, Likeness) market. Founded by Allan Bell, a CPA with deep connections to the NIL ecosystem, the platform allows young athletes to input their performance statistics and receive an AI-generated "Bag Index"—an estimate of their market value based on comparisons to professional athletes in the NBA, NFL, and NCAA.

Beyond valuation, Bag Index provides personalized AI-generated training plans, gamified trading cards, and financial literacy tools to help the next generation of athletes make informed decisions about their careers and earnings.

Challenge:

Building an AI-powered platform in a market that's only two years old, with limited existing tools and documentation. The solution needed to accurately process athletic statistics, generate meaningful comparisons to professional athletes, and present complex financial concepts in an accessible way for young college students.

2. The Problem

Background:

The NIL market opened in 2021, allowing college athletes to profit from their name, image, and likeness for the first time. Suddenly, students who had never earned money from their athletic abilities were signing endorsement deals and making financial decisions with real consequences—but with no tools to understand what they were actually worth.

Allan Bell, a CPA whose son is an NIL athlete, saw this gap firsthand. These young athletes had no way to benchmark their performance against professionals or estimate their market value. They were negotiating deals blind.

Pain Points:

• No existing valuation system: Athletes had no way to estimate their worth based on performance
• Lack of professional benchmarks: No tool to compare college statistics against NBA, NFL, or NCAA professionals
• Financial literacy gap: Young athletes making money for the first time without understanding financial fundamentals
• One-size-fits-all coaching: Generic training plans that don't account for individual statistics and goals
• Fragmented information: Performance data, financial education, and career planning scattered across multiple sources

3. Our Solution

Discovery Process:

Working closely with Allan, we defined a clear phase one scope focused on the core value proposition: a central platform for athletes to track their performance, understand their worth, and plan their development. The client had a well-defined vision, and we delivered everything in the original scope without reduction. As the platform came alive, Allan identified additional features—like AI-based quizzes for financial literacy—which we built as phase two additions.

Core Value Proposition:

A central source of truth for college athletes combining athletic performance tracking with financial literacy education.

Proposed Solution:

• Bag Index Calculation:

  • Athletes input their game statistics (touchdowns, three-pointers, goals, assists, etc.)
  • AI analyzes performance and generates a "Bag Index" score
  • Comparison against profiles of existing NBA, NFL, and NCAA athletes

  • AI Trading Cards:

  • Personalized digital trading cards featuring the athlete's photo and statistics

  • Gamified presentation similar to Pokemon or sports trading cards
  • Shareable assets for social media and personal branding

  • AI-Powered Training Plans:

  • RAG (Retrieval Augmented Generation) system processes athlete's current statistics

  • Generates personalized training recommendations tailored to individual goals
  • Replaces generic one-size-fits-all coaching advice

  • Financial Literacy Quizzes:

  • Administrators input topics, keywords, and descriptions

  • AI automatically generates comprehensive quizzes
  • Educates athletes on financial fundamentals

  • Multi-User Dashboards:

  • Athlete profiles with statistics and progress tracking

  • Parent dashboard for family oversight
  • Coach dashboard for team management
  • Subscription-based access model

Technology Stack:

• Backend: Django, Django REST Framework, Celery (async tasks)
• Frontend: Django HTML Templates, CSS
• Database: PostgreSQL
• Caching: Redis
• Hosting: Heroku
• Payments: Stripe
• AI/ML: OpenAI API (training plans, quizzes), Nano Banana/Google (image generation)
• Sports Data: Sports Radar API
• Code Quality: mypy, ruff, DRF Spectacular

4. Implementation

Platform Development:

• Built a lightweight Django application optimized for performance and reusability
• Implemented subscription model with Stripe integration
• Created role-based dashboards for athletes, parents, and coaches
• Designed responsive web interface accessible on both desktop and mobile

AI Systems Integration:

• RAG Training Plan Generator: Built vector embeddings for athlete documents and statistics, enabling targeted queries to OpenAI for personalized training recommendations. This required careful architecture to avoid simply dumping text into the model—data had to be structured and relevant.

• AI Trading Card Generation: Integrated Nano Banana (Google's image model) to generate consistent, high-quality athlete trading cards. Solved hallucination issues to maintain visual standards across all generated cards.

• Quiz Generation System: Developed automated quiz creation where administrators provide topics and keywords, and the system generates complete assessments using OpenAI.

Sports Radar Integration:

• Integrated Sports Radar API for professional athlete statistics and comparisons
• Overcame limited documentation through extensive research and testing
• Built comprehensive charts and comparison tools for athlete benchmarking

Technical Optimizations:

• Solved N+1 query problems for optimal database performance
• Implemented reusable component architecture across the codebase
• Built caching layer for frequently accessed data
• Added rate limiting and API throttling to control costs
• Designed careful token management to minimize AI API expenses

Technical Challenges Addressed:

1. RAG System Complexity: Building accurate training plan generation required deep understanding of vector embeddings and targeted document retrieval
2. Image Generation Consistency: Taming AI hallucinations in trading card generation to maintain quality standards
3. Undocumented API: Sports Radar's limited documentation required significant research and experimentation
4. Mockup Translation: Converting client's Lovable-generated mockups into production-ready code

5. Results

Product Outcomes:

• Launched fully functional web platform with AI-powered athlete valuation
• Delivered complete phase one scope on budget with no scope reduction
• Client immediately moved forward with phase two development
• Platform now live and serving athletes

Technical Achievements:

• AI-Powered Valuation Engine: Athletes receive personalized Bag Index scores with professional comparisons
• Automated Content Generation: Quiz system creates educational content from simple keyword inputs
• Personalized Training Plans: RAG system delivers tailored recommendations based on individual statistics
• Gamified Experience: Trading cards create shareable, engaging athlete profiles
• Multi-Stakeholder Platform: Unified dashboards for athletes, parents, and coaches

Business Impact:

• First-to-Market Tool: Provides capabilities that didn't exist in this two-year-old market
• Athlete Empowerment: Young athletes can now negotiate NIL deals with data-backed valuations
• Scalable Education: AI quiz generation enables rapid financial literacy content creation
• Personalized Development: Replaces generic coaching with tailored training plans
• Family Involvement: Parent and coach dashboards create support ecosystem around athletes

6. Lessons Learned

Key Takeaways:

• Emerging Markets Require Deep Understanding: The NIL market is only two years old with significant noise. Success required finding signal within that noise to build genuinely valuable features rather than following existing patterns.

• AI at Scale is Now Achievable: Systems we built—RAG-based training plans, automated quiz generation, consistent image generation—wouldn't have been feasible at this quality level even a year ago. The tooling has matured rapidly.

• Clear Scoping Prevents Problems: Mutual respect and clear phase one definition meant zero scope creep. Additional features (like quizzes) were properly scoped as paid additions rather than scope expansion.

• API Documentation Can't Be Assumed: Sports Radar's limited documentation required significant research investment. Building buffer time for third-party API integration is essential.

• Cost Management in AI Applications: Careful architecture around API token usage, caching, and rate limiting keeps AI-powered features economically viable at scale.

7. Conclusion

Summary:

Bag Index brings data-driven decision making to college athletes navigating the new NIL landscape. By combining athletic performance tracking, AI-powered valuation, personalized training plans, and financial literacy education, the platform gives young athletes the tools they need to understand their worth and plan their futures.

The project demonstrates how AI capabilities—from RAG systems to image generation—can be practically applied to solve real problems in emerging markets, delivered through a lightweight, well-architected Django application.