Automation should not be the first question
Understanding real work—including exceptions, waiting and judgement—must precede decisions about what to automate.
Automation projects often begin with a perfectly understandable question: What could this technology do for us? A demonstration creates enthusiasm. A repetitive task becomes visible. Somebody sees the possibility of saving time. The danger is that the proposed solution arrives before we have properly understood the work. Consider a simple approval process. A request arrives. Somebody checks it. A manager approves it. The outcome is recorded. That looks straightforward enough to automate. Now sit beside the person doing it. Perhaps missing information has to be recovered by email. Perhaps they check another spreadsheet because the main system is occasionally wrong.
Perhaps certain cases are passed to an experienced colleague. Perhaps the manager’s approval is normally routine but occasionally involves a judgement that depends upon information not recorded anywhere. Suddenly the process is not quite the process we thought we were automating. None of this means automation is inappropriate. It means we need to know what the work actually contains before deciding what should be automated. Formal processes tend to show activities. Real work also contains waiting, checking, interpretation, exceptions, informal coordination and judgement. Some of that activity is unnecessary friction.
Some of it is what allows the process to work safely. If we automate too early, we risk encoding the friction while accidentally removing the judgement. The alternative does not require months of analysis. Take a few real examples and follow them. Ask what happened. Ask what people checked, chased, corrected and decided. Ask where information came from and what happened when the standard route did not fit. Then decide what the work needs. Sometimes that will be AI. Sometimes conventional automation. Sometimes redesign. Sometimes the answer will be to stop doing something altogether. Technology gives us an expanding range of interventions. Understanding the work is what gives us a basis for choosing between them.