Learn
Guides to AI agents for sport
Build, operate, evaluate, route, and connect context for sports apps and agents, with practical guides for each part of the work.
Start with the platform overview, then choose the guide closest to your task.
Platform and runtime
- Live sports agent runtimeHow an agent runs against live match state: the six-stage loop, the human control point in each stage, and what a one-shot prompt tool cannot do.
- Sports MCP serverWhat an MCP server for sports data exposes, how it differs from a REST API and an agent skill, and the checklist to run before you let an agent depend on one.
- Sports data for AI agentsHow AI agents consume sports data — MCP servers, agent skills and plain REST — what each format gives up, and how to connect one without an API key.
By sports organization
- AI agents for sportsbooksWhat AI agents for sportsbooks are, how they work, the seven use cases operators deploy, and how to choose between building in-house, using a generic AI tool, or an agent platform.
- AI content automation for broadcastersWhat broadcasters actually automate with AI agents, the production constraints that decide whether it survives contact with a live gallery, and how to scope a first deployment.
- AI fan engagement for clubsWhat AI fan engagement actually automates for a club or league, why first-party data decides the outcome, and a checklist for evaluating a platform before you buy one.
Architecture and evaluation
- Sports data API vs AI agent platformWhat a sports data API hands you, what an agent platform hands you, and the work in between that decides which one you actually need.
- Generic framework vs sports-native platformTwo architecture classes compared: what a general-purpose agent framework gives you, what it leaves you to build for sport, and how to judge the gap.
- Build vs buy sports AI agentsA decision framework for building sports AI agents in-house or buying a platform, including the criteria that decide it and the cases where building is the right call.
- How to evaluate a sports AI agent platformA buyer's checklist you can copy into a request for proposal: eight criteria, what strong evidence looks like, and what a weak answer looks like.