Search, Browse & Discover sat at the center of Nike's direct-to-consumer business — every storefront experience, mobile and web, ran through the ranker. 100M+ MAU. The bar for a result page was brand-grade: not just relevant, but on tone. The existing ranker was rule-driven and tightly tuned, and it had been the right shape for years. Generative AI in the critical path of a consumer experience at this scale had no precedent at Nike, and very few elsewhere — the failure mode of an LLM shipping a wrong answer to a hundred million people was a different kind of risk than a search engine returning a less-relevant SKU.

Above the model question sat an org one. Nike Direct Product was 300 people across product, design, business, and engineering — every team optimizing for its own surface. There was no shared register for who the consumer was, what the priorities were across Search vs. Browse vs. Discover vs. PDP vs. Cart, or where the next year's roadmap money should land. Shipping AI in the critical path of a fragmented org would compound the risk: a model is only as defensible as the team that maintains it.

Two teams to lead, simultaneously. A 5-person Central PM function I owned the rebuild of — the small horizontal team that set practice across the 300-person org. And a 30+ person Search/Browse/Discover team I inherited mid-build, with the AI/LLM ranker already in development and political pressure mounting.

Held the line on the AI work. Took over SBD with the ranker already in flight and a real internal argument running against it: merchandisers and marketers inside Nike were pushing that human-led ranking would outperform an LLM, and the safe institutional answer was to defer. The work the team was doing was the right work; my job was to keep it shippable. The case for AI in the critical path had to be made executive-up, not engineer-down — and the team needed someone to make it.

Sharpened the platform-first sequencing. The architectural direction existed when I arrived — shared ranker infrastructure across the surfaces, componentized global navigation, AI on top of platform stability — but the order needed coherence at the strategy layer. I made the sequencing call: platform first so the AI could land on something stable enough to hold it; nav as platform inside the same effort so per-surface re-implementations stopped; instrumentation, brand quality bar, and on-call all routed through the same backbone. Ranker, navigation, and the surfaces around them stopped being three separate engineering investments and started being one.

Initiated and ran the Digital Summit with the team. First gathering of its kind for Nike Direct Product — a multi-day, custom-built session inside the Serena Williams building where product, design, business, and engineering walked through consumer segments together, in purpose-built rooms, until the org shared a register for who they were building for. The Summit wasn't a kickoff; it was an alignment artifact. Its output became the foundation the AI roadmap and the navigation rebuild were both planned against. Same personas, same priorities, same priorities-by-quarter ledger across all four functions. Without the Summit, the model work would have launched into a fragmented brief.

Co-authored the new PM career ladder for the 300-person Nike Direct Product organization out of the Central PM team during the same window. A company-wide restructure was underway and the ladder we shipped became the durable expression of how product would be practiced inside the org — what was expected at each level, how impact was measured, how product partnered with design and engineering. Four pages of doctrine that outlasted any single shipped feature.

The team shipped the AI/LLM ranker. Production search and ranking models in the critical path of every Nike Direct property. Among the first at Nike to put generative AI into the consumer experience at enterprise scale. The model was the headline; the platform underneath, the alignment artifact above it, and the air cover around it are what made the model shippable.

Double-digit clickthrough and engagement lift on the surfaces the AI ranker touched. Nine-figure incremental demand against the existing rule-based baseline. Maxim Award nomination for shipping generative search at consumer scale.

The Digital Summit format became the reference for how Nike Direct Product aligned downstream work — subsequent quarters used the same room-by-room walkthrough as the foundation for roadmap planning. The componentized navigation rolled out across surfaces. The PM career ladder rolled out across the 300-person org and stayed in place through the restructure.

The lane is the lesson. Platform first, alignment ritual that gives the platform a brief, then the innovation moment lands on top of something stable. Search/Browse/Discover at Nike was the third time I'd protected that sequence at scale. The AI work got the headline; the Summit and the ladder are the parts that compound.

Nike Maxim Award nominee — generative search shipping