diff --git a/feature_engine/featurize.py b/feature_engine/featurize.py
index c9ed1dcd..52421316 100644
--- a/feature_engine/featurize.py
+++ b/feature_engine/featurize.py
@@ -28,7 +28,7 @@
 	# )
 	# feature_builder.featurize(col="message")
 
-	# # Tiny multi-task
+	# Tiny multi-task
 	# tiny_multi_task_feature_builder = FeatureBuilder(
 	# 	input_file_path = "../feature_engine/tpm-data/cleaned_data/test_data/multi_task_TINY.csv",
 	# 	vector_directory = "../feature_engine/tpm-data/vector_data/",
diff --git a/feature_engine/utils/preprocess.py b/feature_engine/utils/preprocess.py
index 663e4fdd..0f8ded8b 100644
--- a/feature_engine/utils/preprocess.py
+++ b/feature_engine/utils/preprocess.py
@@ -25,6 +25,10 @@ def assert_key_columns_present(df):
 	# Assert that key columns are present
 	if {'conversation_num', 'message', 'speaker_nickname'}.issubset(df.columns):
 		print("Confirmed that data has `conversation_num`, `message`, and `speaker_nickname` columns!")
+		# ensure no NA's in essential columns
+		df['message'] = df['message'].fillna('')
+		df['conversation_num'] = df['conversation_num'].fillna(0)
+		df['speaker_nickname'] = df['speaker_nickname'].fillna(0)
 	else:
 		print("One of `conversation_num`, `message`, or `speaker_nickname` is missing! Raising error...")
 		print("Columns available: ")