Is Your AI Stack Actually Worth It? A Simple ROI Framework for Founders

Nearly 60% of small businesses used AI tools in 2025. The adoption question is settled. What founders are now asking is different: is this actually working, in numbers I can point to?

The question most founders skip

Most AI decisions are made on feel, not evidence.

You sign up because it sounds useful. You use it inconsistently. You keep paying because cancelling feels like admitting defeat. The result is an AI stack built on assumption rather than data.

This matters because the costs compound. EUR 18 here. EUR 30 there. EUR 10 for the tool someone on your team added to a project six months ago. Add them across a year and you are looking at a real budget line. If those tools are not saving you meaningful time, they are taking your money and your focus without giving much back.

The fix is not complicated. It is a calculation most people skip because they assume the answer is obvious. It rarely is.

The four-step framework

Run this for any tool, or your whole stack, in under 30 minutes.

The part most calculations miss

AI tools do not run themselves.

You spend time writing prompts, reviewing outputs, correcting errors, and iterating. For some tools and tasks, this overhead is minimal. For others, it significantly reduces the net time saving you calculated in step two.

A founder who saves three hours per week of writing time but spends two hours reviewing, editing, and re-prompting has a net saving of one hour, not three. That changes the ROI considerably.

Before finalising your numbers, add a realistic estimate of the weekly time you spend managing each tool. Subtract this from your gross hours saved. Use the net figure in steps three and four.

A worked example

Here is what this looks like with real numbers.

Consider a nine-person branding agency generating around EUR 55,000 per month in revenue. The founder uses four AI tools across the business.

The stack passes the test. But the analysis also reveals that the image generation tool at EUR 30 per month saves roughly 45 minutes of net time after accounting for curation and editing. At EUR 95 per hour, that generates EUR 71 in value for EUR 30 in cost. Marginal, but positive. With team usage factored in, the picture improves. Without it, there is enough information to make a deliberate decision about whether to keep or cancel it.

That is the point of this framework. Not to cut everything that does not deliver a perfect ratio, but to make deliberate decisions with real information rather than assumptions.

When the numbers do not add up

Sometimes a tool is not worth what you are paying for it. That is useful information.

If the time saved is real but the value is low because freed-up hours are not going into billable or high-impact work, the problem is not the tool. It is how you are using the time. Fix the time allocation first, then reassess.

If the tool is genuinely saving very little time relative to its cost and the management overhead is high, cancel it. Redirect the budget toward something with a clearer return.

If the benefits are qualitative, reduced stress, better thinking, more confidence, acknowledge that. But qualitative benefits do not pay invoices. Weight them reasonably and still run the numbers on the measurable side.