For small and midsize businesses: don't overthink the ROI of AI. It depends on one thing: where you put it.
The rule
For an ROI decision, put the work in one of two buckets.
Cost center. You need the work, but you do not sell it directly. Accounting, support, HR, admin.
Profit center. The work directly produces what customers pay for. Delivery, production, the service itself.
AI pays back differently in each. Check the bucket before you buy the tool.
Cost center: faster is not cheaper
Say you run a 40-person company. AI makes your bookkeeper 3x faster.
You still have one bookkeeper. Full time. Same salary. Savings: zero.
In a cost center, the only ROI is fewer people or fewer dollars on the task. If you cannot cut, or skip the next hire, all you get is a faster team at the same cost.
Profit center: speed sells itself
Say you run a five-person design studio. AI removes the first-draft stage.
If demand is there, the same five people ship 40% more projects. Revenue grows without adding payroll.
One catch for service companies: if a 50-hour job now takes 20, hourly billing cuts the invoice to 20 hours as well. Fixed or value-based pricing lets you keep more of the gain.
Before the next AI purchase: two tests
1. Cost center: will we cut spend or avoid a hire?
2. Profit center: can we sell the extra output at a profitable price?
If neither answer is yes, you bought speed, not ROI.
Where are you buying AI speed without a plan to capture the value?
First published on LinkedIn. Discussion on LinkedIn