How to choose your first business automation
A sensible first automation addresses a repeated, understood task with a clear owner and a manageable cost of failure. Before choosing tools, examine the current process, the exceptions, and the effort needed to keep the new workflow working.
The most frustrating task is worth investigating, but frustration alone does not make it the best first project.
Step 1: Define one piece of work.
Write down what starts it and what counts as finished. “Improve customer service” is broad. “Identify appointment enquiries and route them with the required details” gives you something to examine.
Name the person responsible for the task and the people affected by a change.
Step 2: Observe the current effort.
Review recent examples. Estimate how often the work happens, how much active time it takes, and how much waiting or rework occurs. Mark estimates as estimates.
Distinguish time spent doing a task from elapsed time waiting for a decision. Automation may affect those differently.
Step 3: Look at the exceptions.
What happens when a detail is missing, a customer changes the request, or the usual tool is unavailable? How costly is the wrong action? Can the workflow pause and ask a person?
If most cases are exceptions, the first improvement may be to clarify the process or collect better information.
Step 4: Check readiness.
| Question | Evidence to look for |
|---|---|
| Is the responsibility clear? | A named owner and defined completion |
| Are the inputs available? | Examples of the actual information received |
| Are the rules stable? | Written decisions and known exceptions |
| Can the tools support the work? | Confirmed access and integration path |
| Can a person intervene? | A workable review or handoff route |
| Can the outcome be checked? | A baseline and a measurable result |
Do not combine these into a pretend scientific score. Use them to identify what needs attention and whether two candidate tasks have different risks.
Step 5: Compare benefit with the full effort.
Illustrative arithmetic: If a task happens 40 times a week and takes three minutes of active work each time, that is two hours of current weekly effort. It is not automatically two hours of savings.
Subtract the time still required for review, exceptions, correction, and ongoing administration. Consider setup, connected-service costs, and the work required when business rules change.
Freed capacity is not automatically a cash saving. A useful improvement might instead create more reliable follow-up or reduce interruptions. Name the outcome accurately.
Step 6: Agree a test and a stop condition.
Choose representative normal and unusual cases. Define which actions need approval during the initial period. Decide what result would justify expanding the workflow—and what would cause you to pause it.
Review the result against the starting information. Do not declare success based only on the fact that a tool was installed.
A first-project worksheet
- Task: What work are we examining?
- Trigger and finish: Where does it start and end?
- Owner: Who is responsible?
- Volume and effort: What have we observed or estimated?
- Exceptions: When does the normal path fail?
- Tools and access: What must be connected?
- Human review: Which decisions stay with a person?
- Cost and upkeep: What work continues after setup?
- Success measure: What should improve?
- Pause condition: What would make us stop and review?
Need help comparing the options?
Discovery Studio helps turn the current process and candidate improvements into a prioritized plan with visible assumptions.