Right, this is really what's meant by the phrase "fake it until you make it" -- not that you should make claims that aren't true (even though they may be someday) but that you should simulate automating processes in a similar fashion to your anecdote before actually spending the time and money to automate them, even though up front you're likely to lose money on having humans do what you eventually want computers (or robots, in the case of Amazon warehouses) to do.