Working on campaign strategy for the Nets' 60th Anniversary Season, I couldn't get timely insights to support a proposal I believed in. The team went another direction. Riding the subway home, I hit my Notes app: what if AI could have done this in time?

At Brooklyn Sports & Entertainment (BKSE), the path from data request to finished insight took a minimum of two weeks, and quality varied by who had bandwidth. Data rarely shaped early-stage decisions.

Months later, BKSE launched AI Lab, a company-wide competition to build practical AI solutions tied to real business problems. I assembled a cross-functional team of four and entered at the highest tier, “Advanced Innovations”.

With no coding background, I dove into Claude and vibe-coded our first working HTML prototype. Over nine weeks, we refined this early model into Gideon, an AI-powered Insight Narrative Engine. Gideon connects to first and third-party data sources, runs web enrichment, and generates decision-ready strategic narratives benchmarked against historical patterns and league-wide behavior. Rather than a query tool, it functions as an always-available internal analyst built to produce structured insights in the format a strategist actually needs. We believed Gideon could assist our five data analysts serving 15+ teams handling 20+ requests per month with a 50-70% reduction in time from request to finished narrative. Our process required a human in the loop to redirect Gideon away from potential hallucinations, cementing our solution as a tool, not a replacement.

Our live demo included a prototype pulling three use cases from synthetic data, a microsite, a full pitch deck, and a unique narrative built around Elden Ring's Gideon Ofnir, the All-Knowing, with custom shirts for the Barclays Center pitch floor.

We won! $10K, top prize, judged by an external panel from the NBA, AWS Sports, StellarAlgo, and Dataiku out of 14 competing teams. Our CEO greenlit immediate integration (process is ongoing… I'd love to talk about it!).

What started as a subway note became a funded internal product.