The purpose is capability, not transformation
A first project should prove the approach works in your business, build the monitoring and deployment patterns, and give whoever approves budget a result they can point at.
Those outcomes matter more than the size of the saving, because they make everything afterwards cheaper.
Four criteria
- Happens often — daily or weekly
- Well understood — someone can describe every step
- Low consequence if it gets something wrong
- Wanted by whoever currently does it
Finish it properly
- Error handling, not just the happy path
- Monitoring, including alerting on absence
- Documentation sufficient for someone else
- A named owner
- A measurement, agreed before starting
Three half-finished automations are worth less than one finished properly, because none of them can be trusted and all of them need attention.
Measure the before
Two weeks of recording how long the current process takes and how often it happens. That baseline is the only one you will get and it costs almost nothing.
Without it, you will not be able to demonstrate the result, which makes the second project harder to approve.
Review at ninety days
| Question | Answer from |
|---|---|
| Did it save the time expected? | The baseline and current measurement |
| Is it still running correctly? | Monitoring records |
| Has anyone had to intervene? | The owner |
| What should the next one be? | What this one revealed |