diff --git a/homeassistant/components/tts/__init__.py b/homeassistant/components/tts/__init__.py
index 7ee4922677c..719f8c52e7a 100644
--- a/homeassistant/components/tts/__init__.py
+++ b/homeassistant/components/tts/__init__.py
@@ -11,7 +11,7 @@ from typing import Dict, Optional
 
 from aiohttp import web
 import mutagen
-from mutagen.id3 import ID3FileType, TextFrame as ID3Text
+from mutagen.id3 import ID3, TextFrame as ID3Text
 import voluptuous as vol
 
 from homeassistant.components.http import HomeAssistantView
@@ -468,7 +468,9 @@ class SpeechManager:
         try:
             tts_file = mutagen.File(data_bytes)
             if tts_file is not None:
-                if isinstance(tts_file, ID3FileType):
+                if not tts_file.tags:
+                    tts_file.add_tags()
+                if isinstance(tts_file.tags, ID3):
                     tts_file["artist"] = ID3Text(encoding=3, text=artist)
                     tts_file["album"] = ID3Text(encoding=3, text=album)
                     tts_file["title"] = ID3Text(encoding=3, text=message)
diff --git a/tests/components/tts/test_init.py b/tests/components/tts/test_init.py
index d5f099175b1..d66efec8c64 100644
--- a/tests/components/tts/test_init.py
+++ b/tests/components/tts/test_init.py
@@ -73,7 +73,7 @@ def empty_cache_dir(tmp_path, mock_init_cache_dir, mock_get_cache_files, request
     assert False
 
 
-@pytest.fixture(autouse=True)
+@pytest.fixture()
 def mutagen_mock():
     """Mock writing tags."""
     with patch(
@@ -480,7 +480,7 @@ async def test_setup_component_and_test_service_with_receive_voice(
         "42f18378fd4393d18c8dd11d03fa9563c1e54491_en_-_demo.mp3",
         demo_data,
         demo_provider,
-        "AI person is in front of your door.",
+        "There is someone at the door.",
         "en",
         None,
     )
@@ -716,3 +716,24 @@ async def test_setup_component_and_web_get_url_bad_config(hass, hass_client):
 
     req = await client.post(url, json=data)
     assert req.status == 400
+
+
+async def test_tags_with_wave(hass, demo_provider):
+    """Set up the demo platform and call service and receive voice."""
+
+    # below data represents an empty wav file
+    demo_data = bytes.fromhex(
+        "52 49 46 46 24 00 00 00 57 41 56 45 66 6d 74 20 10 00 00 00 01 00 02 00"
+        + "22 56 00 00 88 58 01 00 04 00 10 00 64 61 74 61 00 00 00 00"
+    )
+
+    tagged_data = tts.SpeechManager.write_tags(
+        "42f18378fd4393d18c8dd11d03fa9563c1e54491_en_-_demo.wav",
+        demo_data,
+        demo_provider,
+        "AI person is in front of your door.",
+        "en",
+        None,
+    )
+
+    assert tagged_data != demo_data