If you concatenate two separate images together, Kontext can merge the subjects in the same scene. The problem is, if you prompt "too much", they both lose likeness.