A client recently proposed a project that involved a number of complex automation processes for on-boarding new clients. It would have required three months of planning and two months of coding - just to see it run for the first time.
I suggested they start by automating just three key process steps. Develop it, launch it, and use it.
The simple act of automating three process points began to pay small efficiency dividends immediately. The team had a little extra time; they could breath a little and ponder the next automation steps. They reinvested their newly discovered extra resources in the next automation feature and so on.
The result? An automation golden goose.
A small initial automation investment paid [always increasing] operational dividends with each iteration.
This [agile] approach is self-expanding once the process is primed.
They created a better system because they were able to shape the outcome while in flight.