Orchestration

Also known as: AI Orchestration, Workflow Orchestration

In one sentence

The process of coordinating multiple AI components—model calls, tool integrations, data retrieval, and decision logic—into a coherent workflow that accomplishes complex multi-step tasks.

Explain like I'm 12

Like being the conductor of an orchestra—you don't play every instrument yourself, but you make sure the drums, violins, and trumpets all come in at the right time to create a song that sounds great together.

In context

Real-world AI applications rarely involve a single API call. An AI research assistant might need to search the web, retrieve documents from a database, summarise findings, and draft a report—each step requiring different tools and model calls. Orchestration frameworks like LangChain and LlamaIndex handle this complexity by providing standardised ways to chain operations together, manage state between steps, handle errors gracefully, and route tasks to the right tool or model.