diff --git a/python/samples/core_async_human_in_the_loop/.gitignore b/python/samples/core_async_human_in_the_loop/.gitignore
new file mode 100644
index 000000000..f228262f0
--- /dev/null
+++ b/python/samples/core_async_human_in_the_loop/.gitignore
@@ -0,0 +1 @@
+model_config.json
\ No newline at end of file
diff --git a/python/samples/core_async_human_in_the_loop/main.py b/python/samples/core_async_human_in_the_loop/main.py
index 0c4f3adb4..ccae231b2 100644
--- a/python/samples/core_async_human_in_the_loop/main.py
+++ b/python/samples/core_async_human_in_the_loop/main.py
@@ -28,7 +28,7 @@ import datetime
 import json
 from concurrent.futures import ThreadPoolExecutor
 from dataclasses import dataclass
-from typing import Any, Mapping, Optional
+from typing import Any, Dict, Mapping, Optional
 
 from autogen_core import (
     CancellationToken,
@@ -50,11 +50,10 @@ from autogen_core.models import (
 )
 from autogen_core.tools import BaseTool
 from pydantic import BaseModel, Field
-from utils import get_chat_completion_client_from_envs
 
 
 @dataclass
-class TextMessage(BaseModel):
+class TextMessage:
     source: str
     content: str
 
@@ -113,7 +112,7 @@ class SlowUserProxyAgent(RoutedAgent):
 
     async def save_state(self) -> Mapping[str, Any]:
         state_to_save = {
-            "memory": self._model_context.save_state(),
+            "memory": await self._model_context.save_state(),
         }
         return state_to_save
 
@@ -251,7 +250,7 @@ class TerminationHandler(DefaultInterventionHandler):
         return self.terminateMessage.content
 
 
-async def main(latest_user_input: Optional[str] = None) -> None | str:
+async def main(model_config: Dict[str, Any], latest_user_input: Optional[str] = None) -> None | str:
     """
     Asynchronous function that serves as the entry point of the program.
     This function initializes the necessary components for the program and registers the user and scheduling assistant agents.
@@ -268,6 +267,8 @@ async def main(latest_user_input: Optional[str] = None) -> None | str:
     """
     global state_persister
 
+    model_client = ChatCompletionClient.load_component(model_config)
+
     termination_handler = TerminationHandler()
     needs_user_input_handler = NeedsUserInputHandler()
     runtime = SingleThreadedAgentRuntime(intervention_handlers=[needs_user_input_handler, termination_handler])
@@ -283,11 +284,12 @@ async def main(latest_user_input: Optional[str] = None) -> None | str:
         lambda: SchedulingAssistantAgent(
             "SchedulingAssistant",
             description="AI that helps you schedule meetings",
-            model_client=get_chat_completion_client_from_envs(model="gpt-4o-mini"),
+            model_client=model_client,
             initial_message=initial_schedule_assistant_message,
         ),
     )
 
+    runtime_initiation_message: UserTextMessage | AssistantTextMessage
     if latest_user_input is not None:
         runtime_initiation_message = UserTextMessage(content=latest_user_input, source="User")
     else:
@@ -330,6 +332,9 @@ if __name__ == "__main__":
     # if os.path.exists("state.json"):
     #     os.remove("state.json")
 
+    with open("model_config.json") as f:
+        model_config = json.load(f)
+
     def get_user_input(question_for_user: str):
         print("--------------------------QUESTION_FOR_USER--------------------------")
         print(question_for_user)
@@ -342,7 +347,7 @@ if __name__ == "__main__":
             user_input = get_user_input(question_for_user)
         else:
             user_input = None
-        user_input_needed = await main(user_input)
+        user_input_needed = await main(model_config, user_input)
         if user_input_needed:
             await run_main(user_input_needed)
 
diff --git a/python/samples/core_async_human_in_the_loop/model_config_template.json b/python/samples/core_async_human_in_the_loop/model_config_template.json
new file mode 100644
index 000000000..bfa691390
--- /dev/null
+++ b/python/samples/core_async_human_in_the_loop/model_config_template.json
@@ -0,0 +1,38 @@
+// Use Azure OpenAI with AD token provider.
+// {
+//     "provider": "AzureOpenAIChatCompletionClient",
+//     "config": {
+//         "model": "gpt-4o-2024-05-13",
+//         "azure_endpoint": "https://{your-custom-endpoint}.openai.azure.com/",
+//         "azure_deployment": "{your-azure-deployment}",
+//         "api_version": "2024-06-01",
+//         "azure_ad_token_provider": {
+//             "provider": "autogen_ext.models.openai.AzureTokenProvider",
+//             "config": {
+//                 "provider_kind": "DefaultAzureCredential",
+//                 "scopes": [
+//                     "https://cognitiveservices.azure.com/.default"
+//                 ]
+//             }
+//         }
+//     }
+// }
+// Use Azure Open AI with key
+// {
+//     "provider": "AzureOpenAIChatCompletionClient",
+//     "config": {
+//         "model": "gpt-4o-2024-05-13",
+//         "azure_endpoint": "https://{your-custom-endpoint}.openai.azure.com/",
+//         "azure_deployment": "{your-azure-deployment}",
+//         "api_version": "2024-06-01",
+//         "api_key": "REPLACE_WITH_YOUR_API_KEY"
+//     }
+// }
+// Use Open AI with key
+{
+    "provider": "OpenAIChatCompletionClient",
+    "config": {
+        "model": "gpt-4o-2024-05-13",
+        "api_key": "REPLACE_WITH_YOUR_API_KEY"
+    }
+}
\ No newline at end of file