Add error handling for when LLM provide API calls fall

[TravisCouture]

Occasionally, preflight checks may not catch potential issues ahead of time so we should be able to graceful fail and inform the front end when this occurs.