Business Cases
Part of Business process automation strategy
Choosing a process that is worth automating
Screen automation candidates for value, repeatability, input quality, exceptions, feasibility and effects on other teams.
Test business value and readiness together when choosing a process worth automating. A frequent process with unclear rules may need repair first. A well-defined process may still be a poor priority if improving it would make little difference to the people who use its result.
Build a shortlist from actual problems
Ask process owners and staff for work that repeatedly waits, returns for correction or consumes avoidable effort. Include a customer-facing or downstream team where the result affects them.
For each suggestion, describe the trigger, intended result and problem in one sentence. “Automate admin” is too broad; “reduce the time between a complete request and an approved response” identifies a boundary that can be examined.
Use existing records where available to understand how often the work occurs and what happens to it. A count of items alone says little about the effort, error or consequence involved. Do not assume that the busiest process is the best first candidate.
Apply a value and readiness screen
Discuss every candidate against the same questions. Record evidence and uncertainty instead of assigning a precise score to guesses.
| Screen | Ask | Pause if |
|---|---|---|
| Business value | Would a better result matter to customers, staff or an important business objective? | The benefit is only a vague promise to save time. |
| Repeatability | Do similar inputs follow a route the owner can explain? | Each case requires a different, undocumented judgement. |
| Input quality | Is the information needed for the next action usually available and identifiable? | Staff routinely reconstruct it from messages or memory. |
| Exceptions | Can unusual cases be recognised and sent to an authorised person? | The automation would have to guess or hide a disagreement. |
| Feasibility | Can the proposed route work with the organisation's systems and access? | A critical dependency has no owner or workable path. |
| Wider effect | Would the change help the full process? | It mainly transfers correction work to another team. |
A pause is a preparation task, not necessarily a rejection. Clarifying an intake field or approval rule may make a candidate suitable later. A high-value process without a reliable exception route may be unsuitable for an initial release.
Automation Candidate Screening Criteria
- Business valueWould a better result matter to customers, staff or an important business objective?
- RepeatabilityDo similar inputs follow a route the owner can explain?
- Input qualityIs the information needed for the next action usually available and identifiable?
- ExceptionsCan unusual cases be recognised and sent to an authorised person?
- FeasibilityCan the proposed route work with the organisation's systems and access?
- Wider effectWould the change help the full process?
Pre-Automation Readiness Checklist
- Business value is clear and measurableThe improvement impacts customers, staff or a key business goal like reducing GST processing time.
- Process is repeatable and documentedA consistent path exists for similar inputs; no ad-hoc decisions required.
- Input data is reliable and accessibleRequired fields (e.g., ABN, invoice number) are typically complete and machine-readable.
- Exceptions are identifiable and manageableUnusual cases can be flagged and routed to a human reviewer without automation guesswork.
- Systems and access allow integrationExisting tools (e.g., ATO portals, accounting software) support secure data flow.
- Change won’t shift problems elsewhereImprovement benefits the whole process, not just one team (e.g., not moving errors to superannuation claims).
Compare the improvement options
For the strongest candidates, ask what happens if the team removes an unnecessary step, standardises the normal route or gives staff a better template. Such changes may deliver enough value without process automation. Where repeated routing, checking or handoffs remain, specify the segment that technology could support.
Suppose customer requests are often returned because the initial description is incomplete. An automated handoff would send incomplete work faster. A better intake question may address the cause.
If valid requests still sit unassigned, routing them could then be a useful automation candidate. This is a hypothetical decision sequence, not an observed result.
Select a first scope and a stopping rule
Pick a candidate with a meaningful outcome, a named owner and a normal route that can be described. State which cases are included, who handles exceptions and what result will be verified. Route unusual or high-consequence cases to an authorised person, and assess the controls needed before widening the automation.
Before delivery, record current performance and a measure for the intended result. Choose a guardrail such as corrections or unresolved cases. Decide what evidence would cause the team to expand, change or stop the proposed automation. The first release then produces a useful decision, even if it shows that simplification was the better choice.
Steps to Select a Process Worth Automating
- Pick a candidate with a meaningful outcomeEnsure it has a named owner and a normal route that can be described.
- Define scope and exception handlingState which cases are included, who handles exceptions, and what result will be verified.
- Set controls and guardrailsRoute unusual or high-consequence cases to an authorised person; assess controls before widening automation.
- Record baseline performanceBefore delivery, document current performance and a measure for the intended result.
- Define stopping rulesDecide what evidence would cause the team to expand, change or stop the automation.
