Mohamed Yousuf A

Human Oversight in LLM-Powered Workflow Automation: A Focused Narrative Review of Hallucination Mitigation, Confidence-Based Routing, and Operational Trade-offs

This paper presents a focused narrative review of human oversight in large language model (LLM)-powered workflow automation. It examines the challenges posed by hallucinated or unsupported model outputs and discusses approaches including retrieval-augmented generation, uncertaint...