Case Study: My First AI Experiment
Background
For years, my coworkers and I have bonded over fantasy football through the Sleeper app. Beyond the friendly competition, I always appreciated Sleeper’s intuitive interface and the little UX details that made managing my team a joy.
When I discovered they offered a public API, my mind immediately began racing. As both a sports enthusiast and a data junkie, I saw an opportunity to merge two passions—sports and analytics—into a personal project.
But like many developers with side-project dreams, I never “found the time.”
The Shift in Mindset
Recently, while exploring new career opportunities, I realized I wanted to lean more into product ownership. After 15+ years as a developer—with Agile and Scrum sprinkled in over the last decade—it felt like the right time to transition into a more strategic, product-focused role.
Around that same time, I had a conversation with a coworker about the power of AI in accelerating software development. I decided to put that theory to the test—treating this as both a passion project and a live case study in product creation with AI.
The Idea
I wanted to create an Awake Fantasy Draft Analyzer App—a tool that not only consumed Sleeper’s public API but also:
- Analyzed live and historical fantasy league data
- Produced customized insights based on draft strategy, league format, and player availability
- Adapted analytics dynamically to changing circumstances (injuries, trades, waiver wire movements)
In short, I wanted an app that could do more than show me data—it needed to interpret it like a seasoned fantasy GM.
The Process
Instead of coding from scratch, I decided to flex my product owner muscles. My role was to:
- Define end-to-end product specifications—clear, detailed requirements, architecture diagrams, and data flow.
- Leverage Claude Code as my AI development partner to turn those specifications into working code.
- Iterate rapidly with AI as the “developer” while I played the role of “PO + Architect.”
I designed the app’s logic to:
- Pull data from Sleeper’s endpoints
- Normalize and enrich the datasets for more meaningful comparisons
- Generate complex, scenario-based insights
Execution Timeline
Day 1–2:
- Drafted detailed specifications, use cases, and data models.
- Fed Claude precise prompts to build API integrations and core analytics functions.
- Got a working baseline app within 48 hours—basic but functional.
Day 3–5:
- Added personal touches and fine-tuned logic.
- Customized UI elements for faster decision-making.
- Tested against real league data for accuracy.
Day 6–7:
- Fixed edge-case bugs (bye weeks, unexpected injuries).
- Optimized performance and finalized the V1 MVP.
The Result
By the end of the week, I had a working Fantasy Draft Analyzer—my first-ever AI-assisted product build. The speed and efficiency of AI let me focus on vision, product strategy, and refinement, while the heavy lifting of initial coding was accelerated dramatically.
I felt like Tony Stark creating new toys with Jarvis—designing, orchestrating, and iterating in real-time.
Key Takeaways
- AI is powerful, but it’s not magic — it needs guidance, structure, and a clear vision to deliver meaningful results.
- Experience matters — knowing the underlying engineering principles made me a far better “AI product manager.”
- I have a product mind — this project reaffirmed my ability to see products not just as features, but as systems of interconnected building blocks.
- Rapid prototyping is a game changer — with AI, you can go from concept to MVP in days, not months.
Final Thoughts
The Awake Fantasy Draft Analyzer is more than just a fun side project—it’s proof that AI can be a force multiplier for experienced professionals. The right mix of domain expertise, product vision, and technical understanding can turn a long-shelved idea into a live application in record time.
For me, this was the start of something bigger: using AI not just as a coding shortcut, but as a collaborative partner in product innovation.