Choosing your first automation so the second one gets funded

The highest-return process is not always the right one to build first. The first automation has a second job: convincing the people who control next year's budget that this works. Pick one that fails at that job and the roadmap dies regardless of the hours it saved.

Four things to weigh

• Hours returned — real, measured, and attributable to a named team, not spread thinly across everyone.
• Visibility — someone senior should notice the difference without being told to look.
• Low blast radius — if it fails on day one, the fallback is the old manual process, not a stopped production line.
• A willing owner — one person on the client side who wants it to work and will chase colleagues to use it.

The pattern that works

A process that eats a specific, complaining team's week, is understood well enough to describe in a paragraph, and produces an artefact someone senior already looks at. Approval turnaround times are a good example: the improvement shows up in a number leadership already tracks.

The pattern that doesn't

The biggest, most tangled process, chosen because it's the biggest. It takes three months, it involves four departments who each want something different, and by the time it lands the sponsor has moved on. Build it third, when you have credit in the bank.

Measure the before, honestly, in writing, before you build. Nobody remembers how long it used to take, and without a baseline your success is a matter of opinion.