diff --git a/services/ai_orchestrator_service/voice.py b/services/ai_orchestrator_service/voice.py
index 04654fd..806233d 100644
--- a/services/ai_orchestrator_service/voice.py
+++ b/services/ai_orchestrator_service/voice.py
@@ -375,8 +375,17 @@ def _extract_name_candidate(text: str | None, language: str) -> tuple[str | None
 
     candidate = _normalize_name_candidate(raw)
     if candidate:
+        lower_cand = candidate.lower()
+        stopwords = {
+            "здравствуйте", "привет", "алло", "да", "нет", "добрый", "день",
+            "саламатсыз", "сәлеметсіз", "ба", "бе", "ау", "слышно", "интернет", "не", "работает", "вопрос", "у", "меня"
+        }
+        cand_words = set(lower_cand.split())
+        if cand_words.issubset(stopwords) or len(lower_cand) < 2:
+            return None, False
+
         word_count = len(re.findall(r"[^\W\d_]+(?:[-'][^\W\d_]+)*", raw, flags=re.UNICODE))
-        if word_count <= 3 and not _name_followup_needed(raw):
+        if word_count <= 2 and not _name_followup_needed(raw):
             return candidate, False
         return candidate, True
     return None, False