A compound AI architecture is a system design that combines models with supporting components such as retrieval systems, databases, tools, APIs, workflows, and guardrails. The aim is to deliver behavior that is more context-aware, reliable, and operationally useful than a standalone model.
Compound AI architectures are combinations of one or more models and other components such as databases, APIs, and guardrails that work together to deliver robust, context-aware behavior.