Depending on the use, perhaps, although I think the people it's aimed at would probably have payback in a few weeks.
I did for a while work for Honda in their plant in Swindon. While I was there we spent (I forget the exact figure, but 100k rings a bell) a vast sum of money on a new section of production line to accurately apply sealer to a part before it was attached to the body.
The previous method used a robot to apply the sealer, then a human to check it and apply it to the body. We had several failures per shift - and it was only found at the end of the line during the final checks before the car left the factory. The failures at this stage were enormously expensive as it normally meant another trip through the paint section once it had been repaired.
So using people is OK, but in a critical environment you need to include the cost of mistakes made by human operators.
In the Honda example, we had payback within a month.
