AI projects attract wishful accounting like nothing I have seen since the ERP era. Here is how to measure honestly, drawn from my performance measurement work.
Rule one. Baseline before you build. The single most common failure is starting the project, then trying to reconstruct what the process used to cost. Memory flatters. Before any implementation, measure the current state for two or three cycles. Hours spent, error rates, cycle time, rework, whatever the process actually costs. Thirty minutes of measurement now is the difference between evidence and anecdote later.
Rule two. Count the whole cost. Subscriptions are the visible sliver. The real investment includes setup and integration time, training, the learning-curve dip, and ongoing oversight. Projects that look brilliant against subscription cost alone look merely good against full cost, and merely good is fine, if it is true.
Rule three. Separate three kinds of return, because they behave differently. Time returned. Hours no longer spent, which only become money when redeployed to billable or valuable work, so track the redeployment, not just the saving. Quality returned. Fewer errors, less rework, faster response, measurable in the same units as your baseline. Capacity returned. Things you now do that were previously impossible, more proposals submitted, faster follow-up, analysis you never had time for. The third category is usually the largest and the least measured.
Rule four. Review at ninety days with the numbers, not the demo. Kill or fix what underperforms. Scale what works. An organization that visibly kills a weak AI project earns more internal trust for the next one than any success announcement.
The discipline is boring, which is the point. Boring measurement is what separates organizations compounding real gains from organizations collecting subscriptions.
Mel Meier, DBA is an applied social scientist and the founder of iB Consulting USA, a woman owned, WOSB certified advisory firm. She serves boards, associations, and executive teams through board advisory, organizational diagnostics, speaking engagements in the United States and internationally, and organizational assessments.
