The number changed with every click, and they had to put their signature on it.
A score that doesn't change behind your back and that you can defend line by line in front of whoever has to sign.
Same input. A different score every time.
They had tried to speed up evaluations with an AI-based scoring tool. On paper it was magic: answer the questionnaire, out comes the score. In reality it was a time bomb. The very same input, regenerated ten minutes later, gave a different grade. A client in a meeting would ask “where does this 68 come from?” and the consultant couldn't answer: the model didn't explain the calculation, and the next time the 68 had become 74.
For a boutique that lives on technical credibility, this is not a nuisance: it is poison. The cost was not the software: it was the trust burned in front of the client, the hours lost “checking by hand” an output that was supposed to save you time, and the risk that an evaluation would end up in a binding dossier without a methodology that could withstand a challenge.
Same questionnaire, same input, regenerated multiple times: the score oscillated with no stated reason, from 68 to 74 to 61 to 70. A consultant cannot defend in a meeting a number that not even they can explain.
Defensible or fast. Never together.
They had already tried two paths. Scoring by hand on a spreadsheet: defensible but very slow, and each consultant weighed the items in their own way, so two evaluations of the same company diverged.
And the AI tool: fast but not reproducible, not explainable, and with the fatal flaw of inventing a score even when the data was incomplete, exactly the case in which a professional must say “here I'm missing elements”, not fire off a figure.
The score that doesn't change its mind behind your back.
Five elements, from the questionnaire to the final number, that make the scoring repeatable and explainable in every part.
The questionnaire becomes a 0-100 score computed with fixed weighted formulas: same input, same output, always.
Weighted sub-scores per layer, so the consultant sees which areas the grade comes from.
They translate the number into a verdict readable by the client.
Where a layer is missing, the system declares it, it doesn't make it up.
To the question “where does this number come from?” the answer is on the screen.
Same input. Same output. Always.
Evaluation time reduced by about 60% compared to the manual spreadsheet. Objections in meetings, which used to break trust, now close in thirty seconds by opening the detail. Zero evaluations signed on incomplete data without it being declared.
The exact same input returns the exact same score, always: 100% reproducibility. And the number breaks down into the layers that compose it, so the consultant always knows where each point comes from.
What changes, measured.
Compared to the manual spreadsheet.
Same input, same output, always.
Zero evaluations signed on incomplete data without declaring it.
A score that doesn't change behind your back and that you can defend line by line in front of whoever has to sign.
How many hours do you lose re-checking a score that should already stand on its own? The ROIometro puts it in euros.