Partner Marketing · AI Systems
Ran partner marketing for JAY-Z 30 London, Universal Music, Crypto.com and Tria on AI systems I built.
- 13,072,897 flight itineraries in the demand engine
- 3.17B tokens through my 60-model router
- 322,761 travel market records 24 sources one engine
- 59,609 competitor posts + ads harvested and scored
- 4.5X travel click benchmark crypto.com travel
- $120,000 prize pool launched tria grand prix
- 2M+ merchants in network travel retail and vip
By the Numbers
Crypto.com Travel
The travel click benchmark. Crypto.com Travel members clicked through at 10.9%, against a 2.4% average.
I built the Crypto.com Travel lifecycle: six emails from welcome to level-up, each cut for US, UK and EEA rules.
I sent the welcome myself in six waves. UK members clicked at 14.5%, EEA members at 13.8%.
JAY-Z 30 London
London hotels priced and paired with eight VIP ticket packages.
I ran the JAY-Z 30 London margin desk through the final show, with two live rate pullers reconciled every day.
I wrote and designed the fan email and the launch cards, and animated the Front Row cover with AI.
Tria Grand Prix
Prize pool across nine rounds, plus two Ferrari Paddock Club seats at Suzuka, launched on my films.
I made the Tria Grand Prix launch films with generative video and code-rendered motion: the launch, the Ferrari Paddock Club prize and the round boards, plus the competition teaser.
I shipped every film in square, vertical and landscape for the LinkedIn, X and Instagram launch.
UM Pulse
Launch films: kinetic type in three formats, plus an AI video pack.
I ran comarketing and launch for UM Pulse, Universal Music's fan travel app: the launch pack, the lifecycle emails and the films.
I set the offer in kinetic type and turned concert stills into motion with AI, for the app and socials.
Deal News
Deal-news films: eight Departures boards in three formats, plus the combined cuts.
I designed the Departures board series for BookIt and Superlogic's deal news: 27 crypto partners, the Spree acquisition and Animoca Brands' investment.
I cut Lin Dai's FINTECH.TV segment into captioned clips for every channel.
Bookit Beyond
Member value a year: eight benefits in one membership I designed, priced and modeled.
I designed Bookit Beyond end to end: lounges, hotel credit, cruise credit, live events, experience drops, an AI concierge and dining rewards in one membership.
I built a unit-economics model in code behind every tier and cut the deck for each client.
AI Systems
Agents in one research run: 185 travel data sources mapped, 91 checked by skeptic agents, 1,440 tool calls.
- 01 Sources→
- X, TikTok, Instagram
- Google and Meta ad libraries
- Hotel and cruise rates
- Flight itineraries
- Partner calls
- 02 Engine→
- Self-hosted model router
- Per-partner corpus
- Fail-closed quality gates
- 03 Agents→
- Partner research
- Pricing and margin
- Creative and film
- Research to deck
- 04 Shipped
- Pitch decks in an afternoon
- Lifecycle emails
- Launch films
- Battlecards and dossiers
I built the AI layer the marketing function ran on: multi-agent pipelines for research, creative, pricing and decks, wired into Google Workspace, Slack, Fathom and Linear through MCP servers.
My agents read each partner end to end, from site and app code to ad libraries and social, and kept a living corpus per account. I took research to a print-ready pitch with live revenue models in an afternoon.