diff --git a/keybert/llm/_litellm.py b/keybert/llm/_litellm.py
index f0e55469..7063a93f 100644
--- a/keybert/llm/_litellm.py
+++ b/keybert/llm/_litellm.py
@@ -68,6 +68,7 @@ class LiteLLM(BaseLLM):
     def __init__(self,
                  model: str = "gpt-3.5-turbo",
                  prompt: str = None,
+                 system_content: str = "You are a helpful assistant.",
                  generator_kwargs: Mapping[str, Any] = {},
                  delay_in_seconds: float = None,
                  verbose: bool = False
@@ -79,6 +80,7 @@ def __init__(self,
         else:
             self.prompt = prompt
 
+        self.system_content = system_content
         self.default_prompt_ = DEFAULT_PROMPT
         self.delay_in_seconds = delay_in_seconds
         self.verbose = verbose
@@ -116,7 +118,7 @@ def extract_keywords(self, documents: List[str], candidate_keywords: List[List[s
 
             # Use a chat model
             messages = [
-                {"role": "system", "content": "You are a helpful assistant."},
+                {"role": "system", "content": self.system_content},
                 {"role": "user", "content": prompt}
             ]
             kwargs = {"model": self.model, "messages": messages, **self.generator_kwargs}
diff --git a/keybert/llm/_openai.py b/keybert/llm/_openai.py
index 5c8c078c..70671afa 100644
--- a/keybert/llm/_openai.py
+++ b/keybert/llm/_openai.py
@@ -114,6 +114,7 @@ def __init__(self,
                  client,
                  model: str = "gpt-3.5-turbo-instruct",
                  prompt: str = None,
+                 system_content: str = "You are a helpful assistant.",
                  generator_kwargs: Mapping[str, Any] = {},
                  delay_in_seconds: float = None,
                  exponential_backoff: bool = False,
@@ -128,6 +129,7 @@ def __init__(self,
         else:
             self.prompt = prompt
 
+        self.system_content = system_content
         self.default_prompt_ = DEFAULT_CHAT_PROMPT if chat else DEFAULT_PROMPT
         self.delay_in_seconds = delay_in_seconds
         self.exponential_backoff = exponential_backoff
@@ -170,7 +172,7 @@ def extract_keywords(self, documents: List[str], candidate_keywords: List[List[s
             # Use a chat model
             if self.chat:
                 messages = [
-                    {"role": "system", "content": "You are a helpful assistant."},
+                    {"role": "system", "content": self.system_content},
                     {"role": "user", "content": prompt}
                 ]
                 kwargs = {"model": self.model, "messages": messages, **self.generator_kwargs}