Lets say if we dope a monovalent atom into a 13 atom Ag cluster to make a 13-electrons system with icosahedron symmetry, will it actually fit in the jellium model?
Well, jellium model hinges on the basic assumption that the electrons delocalize and form the jelly like smeared charge. So, how well something fits into jellium, depends on how delocalized/localized the icosahedral system becomes.
Through the Jellium model you can incorporate the superatom concept and see if your systeme would be stable according to this. By that you can see if your cluster uses "atomic orbital"-ish MOs which will help you with the description of the electronic structure, if that's what's of interest to you.
Here's a very good review of this concept for gold clusters by Hannu Häkkinen: