Is there a way to place a "yet to be built" model into a protocol?

For example, I've a protocol that collects data for a compound and it's responses.  Once there are enough responses I'd like to build an RP Tree or RP Forest model.  I'd then like the user of the protocol, via webport, to be able to access the output from the model without having to edit the results protcol.

I thought this could be achieved using variables to name the model and call it but I can't find a way to get it to work.

Thanks for any replys.