Routing & Automation
Part of Support technology improvement
Auditing repetitive work before adding another support tool
Trace repeated agent actions, identify their causes and decide whether a workflow change, existing feature or new tool is justified.
Audit repetitive work by tracing what agents do for a defined set of enquiries. Record repeated actions, find why they recur and decide whether a process change, an existing feature or a new tool could remove the cause. A frequent action may be a necessary verification step.
Define the work to inspect
Choose an issue type and period, then review cases from the channels that handle it. Include straightforward requests, exceptions, handovers and cases that came back. Keep case references in a restricted working sheet rather than copying customer details into the audit.
For each case, note the customer's goal, the agent's steps, the systems opened and any information entered more than once. Ask what the agent was trying to establish.
A second lookup might result from stale displayed data, a missing identifier or a required check before a sensitive action. Each calls for a different remedy.
| Record | Question |
|---|---|
| Repeated action | What was entered, searched, copied or sent again? |
| Trigger | When in the case did it recur? |
| Cause | Was it missing data, unclear ownership, a policy requirement or a tool limit? |
| Customer effect | Did the customer wait, repeat information or receive another message? |
| Safe alternative | What would shorten the step without losing a necessary check? |
Check the pattern across relevant case types and agents. Ask agents to correct mistaken interpretations of their work.
Key metrics to track during audit
- Number of repeated actions observed
- Count across case types and agents
- Average time per repeated action
- Estimate based on observed duration
- Customer wait time due to repetition
- Time added per case from rework or delays
- Cases requiring manual reassignment
- Indicates unclear ownership or process gaps
Distinguish necessary checks from avoidable work
When an approved answer is sent often, it may suit a maintained template. Zendesk, for example, documents macros that agents apply manually to add comment text or change ticket properties.
A macro may help when the agent still needs to judge whether the answer fits; it does not make an unattended reply appropriate.
Other repetition may begin earlier. If an intake path omits a reference agents need, review that path.
If agents paste a verified fact into several systems, identify which record owns it and whether a dependable connection is available. If they reassign a case to discover its owner, settle ownership before automating the transfer.
Count observed repeated actions and, where possible, the time each takes. Any projection needs a stated case volume and an estimate of how often the action occurs.
Do not count necessary checks as recoverable time. Check whether a faster agent step causes another customer contact later.
Necessary checks vs. avoidable repetitive work
- Necessary Check
- Required verification before sensitive action (e.g., confirming account ownership)
- Avoidable Repetition
- Pasting verified facts into multiple systems due to poor integration
- Redundant Lookup
- Second search caused by stale data or missing identifier
- Process Gap
- Intake path omits reference needed for resolution
Decide whether another tool is needed
For each proposed fix, ask whether the current process or subscribed platform can support it, who would maintain it and how a failure would be noticed. Write the required outcome in plain language: what the agent should be able to see or do, and what the customer should no longer have to repeat.
Record the remaining gap after feasible changes to the existing workflow. Consider a new tool when that gap matters to the service, the proposed setup addresses it and the team can maintain the resulting data flow and ownership.
The audit should end with a specific action and owner, whether or not a purchase follows.
