The boring things first
Most of the useful work I’ve done has been unglamorous.
Most of what I’ve done at AANKA is unglamorous. Matching prices across delivery apps, agreeing one version of a menu, making a report send itself so nobody has to ask for it. None of it is clever.
But you cannot improve a number you cannot see. Almost every interesting problem I’ve been handed turned out to be a boring problem underneath. A figure nobody owned, a process that existed in someone’s head, a price that had drifted on one platform and not the others.
The temptation with AI is to start at the impressive end. I’ve found the opposite works better. Fix the thing that is merely tedious, prove it holds, then reach for something harder.