The automation that saved 10 hours a week (and the one that didn't)
Composite, drawn from patterns across engagements, not one specific client.
Two automations, built for the same business within a month of each other. One saved ten hours a week from day one. The other quietly never got used, and nobody flagged it for three months. The difference between them says more than either success story alone.
The one that worked
Weekly reporting, previously a two-hour manual pull from four systems into one spreadsheet, automated into a single dashboard that updated itself. It worked because the underlying process was already well-defined and stable, same four data sources, same format, every single week, for over a year before automation touched it. Automating a stable, well-understood process is close to a guaranteed win.
The one that didn’t
Automated customer follow-up emails, triggered by a status change in the CRM. It looked similar in shape to the reporting win. It failed because the underlying process wasn’t actually stable, sales reps updated that CRM status inconsistently, sometimes early, sometimes late, sometimes not at all. The automation was reliable. The trigger it depended on wasn’t. Reps quietly went back to manual follow-ups within a few weeks, and nobody reported the automation as broken, because from their side, it simply wasn’t relevant to how they actually worked.
What actually separated the two
Not the technology, both were built competently. The difference was whether the human process feeding the automation was already consistent. Automating a consistent process saves time. Automating an inconsistent one just automates the inconsistency, faster and less visibly.
The lesson, generalized
Before automating anything, check whether the underlying human step is already reliable without the automation. If it isn’t, fix that first, or the automation becomes a thing that technically runs and practically does nothing, the worst outcome, because it looks solved on paper while quietly not being solved at all.