Only if the whole job takes less effort at the same quality. Measure from request to accepted result, including review, corrections and cleanup by the people after the first step.
If the draft is faster, isn’t that the saving?
Not if the time moves to someone else. Ask who gained time, and who spent it.
1. Same job, same standard.
Starting point. Record the current steps, the people involved, the effort and the elapsed time.
Comparison. Use similar cases and the same standard for what counts as acceptable.
Quality. Count errors and unfinished work, not only how much was produced.
Practical check: A before-and-after comparison needs a starting point you can defend.
2. The draft is faster. The job is not.
Fictional numbers, to show the arithmetic:
- Before
- 60 staff-minutes to produce an accepted result.
- After
- 65 staff-minutes: 10 preparing, 5 generating, 20 reviewing, 15 correcting and 15 cleaning up later.
- Difference
- 5 extra minutes. The first draft is not the whole process.
3. Time and money are different things.
Running cost. Include licences, usage, support and ongoing upkeep.
Setup effort. Record it separately, and include it when working out whether it pays back.
Time returned. Say what people do with it. Don’t automatically call it money saved.
Practical check: Better quality or service can matter too. Measure it.
4. Keep, change or stop.
- Keep
- The whole result meets the agreed quality and effort test.
- Change
- A specific problem has a correction you can test.
- Stop
- The value doesn’t justify the effort, cost or risk within the agreed limits.
Practical check: More AI activity is not the outcome.
Read more: How to Run a Small AI Agent Pilot · What HR Tasks Can AI Help With? Five to Try
The numbers are fictional and are not a forecast. This is general guidance, not advice for your situation.