Everybody loves processes, right? They make our lives easy: No need to think about it, just follow the process, tick the boxes and boom, effortless result! But is this a good result? A useful result? A meaningful result?
Process-based working certainly has a strong following, their primary arguments being that processes reduce the time it takes to perform a task and that they also reduce the all-too-present, human biases.
Both valid arguments.
What we need to examine however, is whether the benefits of working in this way, outweigh the drawback of losing the human thought process, which is by definition time-consuming and biased.
Here’s a real-life example illustrating the point: A friend of mine used to work for one of the systemic banks of Greece. She was a credit officer and I still remember her telling me back in the day of rapid expansion of the balance sheets of Greek banks, how her work had been reduced to her inputting the assets of the loan applicant in a computer program and then *boom* if their net liquidation value covered that of the loan, it was automatically approved. Need I tell how well did that go? You know, credit crunch hit, loans became non-performing and even when banks where getting hold of the collaterals, they couldn’t sell them. “But we wanted the loan issuance process to be objective you see. Management said that if we put some thought into it, the system would no longer be objective and bad loans could make into the portfolio” she adds.
Compare and contrast the above to what would have happened if people were allowed to put some thought into the loan issuance process and assess future cash flows and the reasoning behind them. Could they prevent all future NPL cases? Probably not. Could they prevent most of them? I believe they could.
We say in valuation that it’s always good to follow two different approaches, so that the results of one are corroborated by those of the other. I believe it’s the same in process-building: Use automation where possible, but always keep a human in the loop. Corroborate the results of the two approaches. Mistakes will still be made, but they will at least be genuinely within reasonable limits.