The Automation Fallacy
Being able to automate half of a job does not mean you have automated half of the jobs. A thought on where the value of work actually sits.
One sentence that is worth turning over a few times:
Just because you can automate 50 percent of a job doesn't mean you can automate 50 percent of the jobs.
Sounds like hair-splitting. It isn't.
The example: the data analyst
What does a data analyst do two thirds of the time? Pull reports. For someone, from some database, dressed up as an Excel wallpaper, a PDF or a PowerPoint deck.
The actual value sits elsewhere: connecting new data sources and making sure the data quality holds up. Not the pulling.
The pulling can be automated. Which data comes in, and whether the finished report says anything at all, are two entirely different skills. They are still part of data analysis. They just never show up in the schedule, because the pulling ate the schedule.
The mistake in the head
The fallacy goes like this: you measure how much of a job consists of repeatable motions, and convert that one to one into "replaceable". Two thirds pulling reports, so two thirds fewer analysts.
But a job is not the sum of its motions. The repeatable parts are the ones you see, because they are loud and eat time. The rest is quiet: deciding which source to trust, noticing that a number cannot be right, not building the report at all because the question was wrong.
Automate the loud part and the job does not disappear. What disappears is the part that made it boring.
The better question
So the question is never "how much of this can I automate?". The question is: which part is the motion, and which part is the judgement? The first part you hand off gladly. With the second, you had better know what you are doing.