diff --git a/vision/cloud-client/detect/beta_snippets.py b/vision/cloud-client/detect/beta_snippets.py
index 5d792b9138bc..52dd00116c3a 100644
--- a/vision/cloud-client/detect/beta_snippets.py
+++ b/vision/cloud-client/detect/beta_snippets.py
@@ -388,9 +388,9 @@ def async_batch_annotate_images_uri(input_image_uri, output_uri):
             localize_objects_uri(args.uri)
         elif 'handwritten-ocr-uri' in args.command:
             detect_handwritten_ocr_uri(args.uri)
-        elif 'batch-annotate-files' in args.command:
+        elif 'batch-annotate-files-uri' in args.command:
             detect_batch_annotate_files_uri(args.uri)
-        elif 'batch-annotate-images' in args.command:
+        elif 'batch-annotate-images-uri' in args.command:
             async_batch_annotate_images_uri(args.uri, args.output)
     else:
         if 'object-localization' in args.command:
diff --git a/vision/cloud-client/detect/beta_snippets_test.py b/vision/cloud-client/detect/beta_snippets_test.py
index acbb5b7015e2..57f7c375ba74 100644
--- a/vision/cloud-client/detect/beta_snippets_test.py
+++ b/vision/cloud-client/detect/beta_snippets_test.py
@@ -79,5 +79,11 @@ def test_async_batch_annotate_images(capsys):
     gcs_uri = GCS_ROOT + 'landmark/eiffel_tower.jpg'
     beta_snippets.async_batch_annotate_images_uri(gcs_uri, GCS_DESTINATION_URI)
     out, _ = capsys.readouterr()
-    assert 'language_code: "en"' in out
     assert 'description: "Tower"' in out
+
+    from google.cloud import storage
+    storage_client = storage.Client()
+    bucket = storage_client.get_bucket(BUCKET)
+    if len(list(bucket.list_blobs(prefix=OUTPUT_PREFIX))) > 0:
+        for blob in bucket.list_blobs(prefix=OUTPUT_PREFIX):
+            blob.delete()
diff --git a/vision/cloud-client/detect/detect.py b/vision/cloud-client/detect/detect.py
index 616fdec3ee19..83974d15d209 100644
--- a/vision/cloud-client/detect/detect.py
+++ b/vision/cloud-client/detect/detect.py
@@ -715,7 +715,7 @@ def async_detect_document(gcs_source_uri, gcs_destination_uri):
     # Process the first output file from GCS.
     # Since we specified batch_size=2, the first response contains
     # the first two pages of the input file.
-    output = blob_list[1]
+    output = blob_list[0]
 
     json_string = output.download_as_string()
     response = json_format.Parse(