That's a really insightful question! You've hit on some of the biggest practical challenges that often get overlooked in those glossy case studies. From what I've seen, there are a few reasons why these realities get glossed over:

First, sensor calibration drift is a constant battle in any automated system. Sensors naturally degrade over time due to aging, temperature changes, and environmental factors. While some advanced systems use AI for auto-drift compensation, it's still a maintenance headache that requires regular attention.

Network latency is another hidden cost. Those 'lights-out' operations demand ultra-low latency communications with redundant pathways, but real-world network issues can still disrupt operations. Even with 5G and advanced networking, you're dealing with thousands of simultaneous data streams that need perfect coordination.

And you're absolutely right about human expertise - it's still crucial! Most 'lights-out' factories aren't truly human-free. They're more like 'lights-sparse' operations where skilled technicians monitor systems, handle unexpected edge cases, and perform maintenance that automation can't manage. The reality is that automation handles routine tasks well, but humans are still needed for troubleshooting, quality control, and dealing with the unpredictable.

Case studies often focus on the success stories and ROI numbers (20-30% productivity improvements) while downplaying the ongoing maintenance costs and human oversight required. It's the classic gap between marketing hype and operational reality!