Agentic GraphRAG for a $100B Industry: Finding Sports Sponsors with Neo4j
Session track: AI Engineering
Session time:
Session description:
Sports sponsorship is a $100B industry, yet after interviewing 200+ pro-sports executives across Europe, Louis and his co-founders, Martin and Clement, found that not one had a structured data process. Deals still close on Excel files and gut feeling, while TV-rights revenue collapses and clubs scramble for sponsors. At LeadSponsor, his team rebuilt that workflow on Neo4j, and in this session, Louis will show you how.
You will learn how to design a knowledge graph for B2B sales intelligence, modeling sponsorship deals as rich (:Brand)-[:SPONSORS]->(:RightHolder) edges with 25+ deal types, tiered partnership states, and list-valued geographic_scope for country-gated queries. You will see, in real Cypher and real code, how a single LLM router dispatches each user question across five specialized tools: pre-built Cypher templates, vector entity search on Brand and Athlete embeddings, full-text retrieval over chunked news articles, LLM-generated Cypher for multi-hop questions, and a hybrid fallback that fuses graph context, vector hits, and live web search in parallel.
You will also learn how they ingest real-time market signals as Article and ArticleChunk nodes with commercial-sentiment scoring, how fuzzy entity resolution at query time keeps "PSG" mapped to "Paris Saint-Germain" without pre-computed aliases, and how they guard against LLM-generated Cypher that hallucinates template names or sneaks in write keywords.
Expect live Cypher, schema-in-prompt patterns, and honest lessons learned from shipping this agent to production sports clubs.
Speakers
Louis Rodriguez
CTO & Co-founder, LeadSponsor
Louis Rodriguez is the CTO and co-founder of LeadSponsor, where he is building the agentic GraphRAG that finds and closes sponsors for pro sports clubs. He holds a Master's in Econometrics & AI from Toulouse School of Economics and graduated from X-HEC Entrepreneurs (joint program of HEC Paris and École Polytechnique). He won the J.P. Morgan Data for Good Hackathon, and now spends his days modeling a $100B industry into Cypher. He is a Neo4j-first engineer who believes the next decade of B2B sales tooling will be built on knowledge graphs, not on Excel files.
Sascha Cauchon
AI Engineer, Dataxx
I'm a recent engineering graduate from Paris and an AI Engineer at Dataxx, where we(re building an agentic graph-RAG platform for the sports industry. Before this, I completed several data analyst internships across companies of all sizes, sharpening my ability to turn messy data into clear decisions. I'm a curious learner at heart, recently placing 3rd at my first fintech hackathon. Off the clock, I cycle to lose myself in new landscapes and play football for the shared passion it brings.
