Yes and no. If you are looking at the whole set of contextual sound alternations across languages, there is little realistic hope for demanding some degree of phonetic "naturalness" to such alternations: see this dissertation by Chabot. However, if you are looking just at low-level contextual phonetic readjustments e.g. coarticulatory effects, then the change is prima facie suspicious. However, it is also attested (such is the nature of prima facie suspicions). The sibilants [ʃ, ʒ] in Russian cause vowel retraction, because they are actually themselves retracted. Thus the phonetic question becomes, what are the articulatory properties of შ – not prejudicing the question by asserting a specific phonetic transcription of that sibilant. The letter ʃ is used cross-linguistically to refer to a wide range of postalveolar fricatives, some being more palatal but some being velarized or retroflex. ʃ is often but not necessarily palatalized, so it depends on what kind of ʃ exists in Georgian.