What We Deliberately Don't Put a Price On

We could convert employee satisfaction into francs. We could attach neat franc amounts to a better customer experience, an innovation image and secured knowledge, and add them to the ROI. 80 percent would swiftly become 140. We do not do it. Deliberately. And that deliberation is not a sacrifice but one of the most honest statements a provider can make about its own numbers.

Eight levers that are real - and do not appear on the invoice

When AI takes effect in a company, effects arise well beyond the hours saved. We encounter eight of them again and again, and each has a genuine impact:

• Higher employee satisfaction, because monotonous tasks fall away and room opens up for more interesting work.
• Fewer errors, because processes run more cleanly and reproducibly.
• Faster response times, towards customers and internally.
• A better customer experience, which translates into loyalty and referrals.
• An innovation image in the market that helps in tenders and in recruiting.
• Scalability without headcount growth, meaning growth that does not hinge on every new hire.
• Better data security and compliance documentation, often stricter than in the manual process.
• Secured knowledge, because implicit skill migrates into prompts and models instead of walking out of the building with people.

All eight are real. All eight influence the value of an AI investment, sometimes more than the pure time saving. And still we convert not a single one of them into a franc amount. That is not an oversight and not laziness. It is a decision with a reason.

The temptation is great - and that is exactly why it is dangerous

It would be easy to act differently. Monetising qualitative factors is the most popular way to dress up a thin business case. "Lower turnover saves us CHF 150,000 in recruitment costs", "higher satisfaction brings 8 percent more productivity", "the image is worth one new client per quarter" - sentences like these sound assured and push the ROI upwards. They are seductive because they turn a weak number into a strong one without changing anything about the actual project. That very ease is the warning sign.

Why not? Because every monetisation is an assumption

The moment you translate "employee satisfaction plus 12 percent" into "CHF 80,000 a year", you have invented a number. Where do the 80,000 come from? From a study that may or may not fit your company. From a rule of thumb someone once wrote down. From the quiet wish to make the ROI look bigger. Every one of these translations is an additional assumption, and every additional assumption is an additional point of attack.

A CFO who knows the craft will hunt for exactly those spots. They do not ask about the hard numbers - those are traceable. They ask about the soft CHF 80,000 assumption, take it apart in thirty seconds, and suddenly it is not just that one number in question but the credibility of the entire calculation. What was meant to reinforce your business case becomes the weak point that drags your solid, defensible numbers into doubt with it. An inflated ROI holds exactly as long as nobody looks closely - and in the boardroom somebody always looks closely.

80 percent with image is not 120 percent without

The real point is not that qualitative factors are worthless. On the contrary - they are often what carries an investment over the long run. The point is that they change the character of an investment, not its return. And character and return are two different things that should not be pressed into the same number.

An AI project with 80 percent of solid, defensible ROI that also relieves the team, lowers the error rate and sharpens the innovation image is a different investment from one with a bare, optimistically calculated 120 percent and none of those levers. Not categorically better or worse in terms of a single metric - different, with a different profile. Perhaps the 80 percent variant is the smarter one, because it works more sustainably and carries less risk. But that distinction disappears entirely the moment you press everything into the same franc figure. Then 120 stands above 80, full stop, and the more important information has been calculated away.

Qualitative factors belong visibly alongside, not hidden inside

So we deliberately keep them separate: as their own block that enriches the decision instead of flattering the return. The decision-maker sees the hard numbers - ROI, break-even, NPV, IRR - and next to them, clearly marked as qualitative, the eight levers that also take effect. They can weight them as fits their company, instead of having a foreign monetisation assumption imposed on them. That respects their judgement instead of replacing it with an invented number.

Honesty as differentiation

In a market where almost every provider promises the maximum possible ROI and generously smuggles qualitative effects into the calculation, restraint is a statement. Saying "we deliberately leave these eight effects out, because we are not selling you invented numbers" signals something no percentage, however high, can deliver: that the numbers being shown will survive scrutiny.

Trust does not come from the biggest number. It comes from the number that still stands unchanged after the third critical question. And in the long run that is the more persuasive sales argument - especially with the decision-makers who have watched one too many inflated business cases break against reality.

That is exactly how our AI ROI calculator is built: hard numbers that hold up, and qualitative factors as a transparent, separate block alongside - not baked in, but shown.

→ See the report