Your question is Error Handling and Human Checkpoints. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How do you implement robust error handling and human-in-the-loop checkpoints in an autonomous agent at Capital One?