Start by noticing where work repeatedly stalls. For one week, jot down tasks that required rework, switching between apps, or hunting for information. Pick the one that happens often and has a clear finish line. A keyboard shortcut or bookmark helps only when it removes friction from that real task.
Batch similar low-risk work, keep a short list of active priorities, and define what 'done' means before opening a research rabbit hole. Automate a step only after its input, output, and failure cases are clear; otherwise the automation can create faster mistakes. Review the result a week later. Saving time is useful, but reducing avoidable errors and mental context switches is usually the larger win.
Akousa