[ericfr11]

It makes somewhat sense that composing an API call would be easier (for an LLM) than inferring a tool call. It will make it easier to observe.