Measure the time to an accepted result, including preparation, review and handover.
We would be uncomfortable selling you a time saving that ends at the first draft. Your business pays for the whole journey to an accepted result: finding the information, explaining the exception, correcting the output and getting someone to act. A demonstration can make the drafting look almost free while leaving the most expensive work untouched.
Our instinct is to ask where the effort goes after the impressive part. Who checks the result? Who gets interrupted when it is wrong? Who maintains the workflow when the input changes? We owe you a scope that recognises those people and that work.
Measure the whole process
An illustrative team might reduce report drafting from an hour to several minutes, then spend longer checking the new output because the sources are unclear. Another team might achieve a smaller drafting gain but receive a reconciliation that makes review easier. The second workflow could be more valuable even though its demonstration is less dramatic.
A practical evaluation begins with a baseline. Describe the existing process and identify who contributes time to it. Include the interruptions that do not appear in a formal procedure: someone finding the right file, explaining a mapping or checking an unusual entry. These details often reveal where the real burden sits.
Our instinct is to ask where the effort goes after the impressive part.
Then define a suitable comparison. Use a representative task with familiar outputs, including the awkward cases. Keep the quality threshold consistent. Record the effort required to reach that threshold rather than comparing a rough AI draft with a fully reviewed manual report. This creates a fairer picture of what changed.
Compare accepted outputs
We also think released time deserves a practical destination. If the team gains capacity, what would you want them to do with it? Speak to customers sooner, investigate a recurring problem, prepare for next week? We would be careful about turning every saved minute into a claimed cash saving. Capacity becomes useful when you choose how to use it.
Some benefits are difficult to express as immediate savings. A clearer report may support a better discussion. A more accessible forecast may reduce dependence on one person. A reviewable process may make handover easier. Such benefits deserve attention, but they should be described as observations or hypotheses until there is evidence of their effect.
Capacity is particularly easy to misunderstand. Time released from a repetitive task becomes valuable when the business has a useful way to spend it. Ask what the team would do with that attention. Investigating exceptions, speaking to customers or planning ahead may be more meaningful than claiming that every saved minute becomes a cash saving.
An honest business case can make adoption easier. People are less likely to feel pressured by an unrealistic promise, and leaders can distinguish a worthwhile experiment from a larger commitment. The evaluation becomes a learning process, with clear reasons to continue, adjust or stop.
What would have to change for this project to be worth your attention? We would agree that test before building and include the cost of review. A result that gives your team a little more reliable capacity every week may deserve more confidence than a dramatic number on a sales slide.
Examples in this essay are illustrative, rather than accounts of client engagements.