“Automation is just the exercise of power, and unwise automation can do as much damage as wise automation can bring benefit.”
I came across this quote from Ernest Mueller in doing some reading on DevOps, and it stopped me in my tracks. “Automation is the exercise of power.”
This way of stating it - that it’s an exercise of power - is what really caught me.
The idea is that anything that we automate or hard code or build into a routine is an exercise of power - whoever owns and reinforces the beliefs and assumptions gets to define what gets automated. The challenge is that the beliefs and assumptions are rarely articulated in a way that we can look at them and change the system that has been created to support.
If we don’t have a way to challenge the assumptions, they get baked into anything that gets “set in stone.” And then it becomes harder to unwind changes and improvements.
The beauty of systems thinking approaches is that we try to understand the situation before building “solutions.” Small experiments to test an idea; Checking for understanding of the deeper reality (will the solution truly remove the source of the problem, or are we playing whack-a-mole?); Testing the ideas with customers; and other approaches.
One of the ideas from Theory of Constraints that I have mentioned previously are the “Questions for Technology.” These help us understand why the technology (or automation, or even other types of change) is interesting. But it also checks that we understand the rules and workarounds that have been in play even though we don’t have that capability.
What is the power of the new technology?
What current limitation or barrier (that exists today) does the new technology eliminate or vastly reduce?
What policies, norms and behavior patterns are used today to bypass the limitation?
What policies, norms and behavior patterns should be used once the new technology is in place?
In view of the above, what changes/additions to the new technology should be introduced?
Just because we have automated, doesn’t mean that we’ve solved the core problem. But if we automate the wrong thing, it can become harder to see and solve the core problem.