Wilcoxon signed-rank test - Can I use for this case?

Hello guys,

I'm trying test if the results of a measurement variable (obtained with observation) is significantly greater than a other (obtained with random process). For the observed, I have an algorithm that returns a number. For the random I use the Bernoulli process, that returns another number. Both answers give me a number between 0 and 1, that can be similarly interpreted. I have a paired comparison and the data is non-normally distributed. So, for example, I have:

Individual / Observed / Random
A / 0.6 / 0.3
B / 0.7 / 0.1
C / 0.4 / 0.25

My question is if I can use Wilcoxon signed-rank test since I am not applying exactly the same measurement variable for each pair, but they are equivalent (again, both between 0 and 1, that can be similarly interpreted).
If it not possible, which test you advice me to use?
Is there some test that fits in such kind of comparison (observation x random).

thanks in advance