Define the outcome and its measure
Delivered orders, time to resolution, contribution margin, customer experience or actual paid labour cost?
Before AI or process automation
Will automating this task improve the whole process?
A task becomes faster, but the customer still waits. The queue may have moved to approval, missing information or delivery. Before automating, check what determines the total time and whether the proposed change will improve it.
The short answer
Automation is more likely to create measurable value when the task is a material constraint, a significant source of avoidable errors, or a cost that can actually be reduced. Check the end-to-end workflow, exceptions, available downstream capacity and how freed time will be used before treating vendor time-savings estimates as business ROI.
Five questions
01 · Goal
02 · Reality
03 · Capacity
04 · Cost
05 · Operation
Follow the work beyond the automated task
Locate the current queue
Record the real path, including incomplete inputs, exceptional cases and approvals.
Check the next team’s capacity
If the next team cannot absorb more work, overall throughput may stay the same.
Specify how released time will be used
Can it eliminate paid hours, avoid a planned hire or be redeployed to valuable work?
Assign responsibility for exceptions
Define fallbacks, maintenance, data quality, testing and the person accountable for results.
Illustrative capacity example
A faster task can move the queue downstream
Suppose a team can prepare 30 orders per day and delivery can complete 40. Automating administration raises preparation capacity to 70. Without a change to delivery, completed orders cannot exceed 40 per day. The automation may still save effort or improve accuracy, but a promise to deliver 70 orders per day would be unsupported.
Before30Orders prepared/day
After automation70Preparation capacity/day
Delivery limit40Orders fulfilled/day
Illustrative capacities, used to show how a downstream limit affects the result.
A useful decision
A pilot can test the overall effect
Not every automation needs a lengthy study. A low-cost, reversible improvement may need only a simple test. A project that changes sales handovers, data ownership and delivery capacity needs a more careful review.
The aim is to establish which assumption could materially change the decision—and what evidence would test it. If the business case is sound, proceeding is a valid outcome.
Before specifying a system
Measure throughput, cost or service, not task speed alone
The case for automation depends on whether the current delay, error or rework originates in the step proposed for automation. Record the baseline, the exceptions that will remain and the work required on either side of the automated step.
For an example of activity status being mistaken for commercial progress, see the monday CRM board analysis.
What to examine next
Check performance at the next handover
Local speed does not necessarily increase end-to-end capacity.