Human-in-the-Loop Enterprise Automation: Orchestrating Workspace Workflows with LLM Parsing

Enterprise workflow automation frequently faces adoption barriers due to strict governance, budget constraints, and risk aversion toward fully autonomous AI agents. This paper presents a zero-infrastructure, "Human-in-the-Loop" (HITL) architectural framework that combines native spreadsheet engines, automated script triggers, and side-panel Large Language Model (LLM) parsing. By enforcing strict boundary conditions where AI performs contextual extraction while deterministic scripts handle execution, the architecture achieves 100% auditable workflow synchronization across multi-project streams.