Assessment · no account
Do you actually need this? Fourteen questions, and an answer that can be no.
Half the questions are about your storefront, half about your traffic. The first half works out which stages an AI agent would fail on. The second turns that into money rather than a score out of a hundred.
How this stays honest.
Not knowing costs you nothing
Every question offers “not sure”, and that answer is excluded from the total rather than counted as a failure. An assessment that charged for your uncertainty would be inflating its own headline figure.
The thresholds are printed
The result shows the cutoffs that produced it, and the inputs it used. You can disagree with the arithmetic, which is the point — a number you cannot check is not worth having.
It uses the product’s own model
The figure comes from the same calculation tragent runs on a real report, not a separate sales model. A return cannot block an order, so it is counted at zero; a stage blocked by an earlier one is not charged twice.
Nothing you answer is stored or sent anywhere but the calculation. A free scan replaces these assumptions with what your live site actually publishes.