The fastest way to waste an automation budget is to automate a broken process. All you get is the same mess, running faster and harder to fix. Before any tooling, the question is the unglamorous one: what is the work, really?
The map you don’t have
Most teams have never seen their own process drawn end to end — who touches what, where the time goes, where the handoffs stall, which steps exist only because someone left years ago. Until that map exists, every automation decision is a guess.
Automate the right thing, not the loud thing
The step everyone complains about is not always the one worth automating. We score opportunities by impact, effort and risk — and just as importantly, we name the ones that should stay human. Honest subtraction is part of the job:
- High impact, low risk — automate first, measure the return.
- High impact, high risk — automate with guardrails and a human in the loop.
- Low impact — leave it alone, no matter how annoying it feels.
Automating a broken process just gives you a faster broken process.
Prove it, then scale it
Once the map is clear and the targets are chosen, the build is the easy part — and you can actually measure what it returned, because you knew the baseline. That is the order that works: map, prioritize, build, measure. Skipping the first step is why so much automation quietly underdelivers.