Reviewing whether automation moved work to another team: Trace one item from sender to receiver; compare effort, waiting and corrections.; A notification sent or task created does not prove the recipient finished the work.; Record eligible population and active labour minutes by team over comparable periods.
Image: Business Automation Desk

Human Oversight

Part of Automation programme reviews

Reviewing whether automation moved work to another team

Trace cases across a handoff and compare effort, corrections and outcomes in both teams before claiming an automation benefit.

Desk Programme · · 3 min read

To check whether an automation moved work, trace the same business item from the sending team through the receiving team to its final result. Compare active effort, waiting and corrections in both teams before and after the change. A reduction in the sending team's handling time is a local result until the downstream effect is known.

Trace one case from the sending team to the final result

  1. Choose a case type and identify what the automation sends and to whom
  2. State what counts as acceptance and as a useful completed result
  3. Follow the same business item through the receiving team to its final result
  4. Compare active effort, waiting time and corrections in both teams, before and after the change
  5. Describe the net effect before treating a local time saving as a benefit

Define the handoff

Choose a case type and identify what the automation sends, to whom and what the receiving team must do with it. State what counts as acceptance and as a useful completed result. A notification sent or task created does not establish that the recipient finished the work.

Ask both teams to show ordinary, returned and still-open cases. Record who checked an output, supplied missing information or repaired a destination record. Work may move into a shared queue, spreadsheet or specialist's interruptions; include those routes when they are part of finishing the same case.

Evidence to collect on both sides of the handoff

  • Ordinary, returned and still-open cases shown by both teams
  • Who checked an output before it was accepted
  • Who supplied missing information needed to finish the case
  • Who repaired a destination record
  • Shared queues, spreadsheets and specialist interruptions included as part of the route
  • Active labour minutes and waiting time recorded separately
  • The same unit of work, such as one request, used in both teams

Count work on both sides

For comparable periods and case types, record the eligible population and active labour minutes by team. Keep waiting time separate from active effort. Count returned and unresolved cases, and note where a problem was discovered. A team may spend fewer minutes per item while cases wait longer downstream.

Possible finding / What to verify

Sending-team effort fell
Which steps disappeared, and for which cases?
Receiving-team effort rose
Was the work new, transferred or previously unrecorded?
More cases are returned
What information or decision is missing at the handoff?
The final result improved
Is the improvement visible in completed cases despite added checking?

Use the same unit across teams, such as one request, rather than comparing bot runs with people's tasks. Where handling time comes from a sample, show its size, selection and case mix. An observed transfer in a few cases does not establish a programme-wide saving or loss.

Match the finding to what you need to verify

  • Sending-team effort fellVerify which steps disappeared, and for which cases
  • Receiving-team effort roseVerify whether the work is new, transferred or was previously unrecorded
  • More cases are returnedFind the information or decision that is missing at the handoff
  • The final result improvedCheck whether the gain is visible in completed cases despite the added checking

Interpret the transfer

Moved work is not automatically waste. A receiving team may now perform a necessary review that was previously missed. Conversely, an automation may send incomplete items faster and force the recipient to reconstruct information. Ask whether the new work contributes to the intended result, whether the team has authority to do it and whether the handoff supplies what it needs.

Describe the net effect: which team released capacity, which took on work, what happened to elapsed time and quality, and what remains unknown. If the handoff is poor, change the input or route and measure again. If the added check is necessary, include its effort in the realised-value account.

More from Human Oversight