diff --git a/llm_observability_examples.py b/llm_observability_examples.py
index 8328780..9b9a42d 100644
--- a/llm_observability_examples.py
+++ b/llm_observability_examples.py
@@ -31,6 +31,7 @@ def main_sync():
     try:
         basic_openai_call(distinct_id, trace_id, properties)
         # streaming_openai_call(distinct_id, trace_id, properties)
+        non_instrumented_openai_call()
     except Exception as e:
         print("Error during OpenAI call:", str(e))
 
@@ -130,6 +131,12 @@ async def streaming_async_openai_call(distinct_id, trace_id, properties):
     return response
 
 
+def non_instrumented_openai_call():
+    response = openai_client.images.generate(model="dall-e-3", prompt="A cute baby sea otter", n=1, size="1024x1024")
+    print(response)
+    return response
+
+
 # HOW TO RUN:
 # comment out one of these to run the other
 
diff --git a/posthog/ai/__init__.py b/posthog/ai/__init__.py
index e3d4311..6fbccd1 100644
--- a/posthog/ai/__init__.py
+++ b/posthog/ai/__init__.py
@@ -2,4 +2,3 @@
 from .providers.openai.openai_async import AsyncOpenAI
 
 __all__ = ["OpenAI", "AsyncOpenAI"]
-# TODO: add Azure OpenAI wrapper
diff --git a/posthog/ai/providers/openai/openai.py b/posthog/ai/providers/openai/openai.py
index 7bbae23..58accda 100644
--- a/posthog/ai/providers/openai/openai.py
+++ b/posthog/ai/providers/openai/openai.py
@@ -8,7 +8,7 @@
 
 from posthog.client import Client as PostHogClient
 from posthog.ai.utils import (
-    track_usage,
+    call_llm_and_track_usage,
     get_model_params,
 )
 
@@ -33,6 +33,15 @@ def __init__(
         self._openai_client = openai.OpenAI(**openai_config)
         self._posthog_client = posthog_client
 
+    def __getattr__(self, name: str) -> Any:
+        """
+        Expose all attributes of the underlying openai.OpenAI instance except for the 'chat' property,
+        which is replaced with a custom ChatNamespace for usage tracking.
+        """
+        if name == "chat":
+            return self.chat
+        return getattr(self._openai_client, name)
+
     @property
     def chat(self) -> "ChatNamespace":
         return ChatNamespace(self._posthog_client, self._openai_client)
@@ -61,7 +70,7 @@ def create(
         **kwargs: Any,
     ):
         distinct_id = posthog_distinct_id or "anonymous_ai_user"
-        
+
         if kwargs.get("stream", False):
             return self._create_streaming(
                 distinct_id,
@@ -70,11 +79,10 @@ def create(
                 **kwargs,
             )
 
-        
         def call_method(**call_kwargs):
             return self._openai_client.chat.completions.create(**call_kwargs)
 
-        return track_usage(
+        return call_llm_and_track_usage(
             distinct_id,
             self._ph_client,
             posthog_trace_id,
@@ -113,7 +121,9 @@ def generator():
                 end_time = time.time()
                 latency = end_time - start_time
                 output = "".join(accumulated_content)
-                self._capture_streaming_event(distinct_id, posthog_trace_id, posthog_properties, kwargs, usage_stats, latency, output)
+                self._capture_streaming_event(
+                    distinct_id, posthog_trace_id, posthog_properties, kwargs, usage_stats, latency, output
+                )
 
         return generator()
 
@@ -127,7 +137,7 @@ def _capture_streaming_event(
         latency: float,
         output: str,
     ):
-        
+
         event_properties = {
             "$ai_provider": "openai",
             "$ai_model": kwargs.get("model"),
@@ -140,7 +150,7 @@ def _capture_streaming_event(
                         "role": "assistant",
                     }
                 ]
-            }, 
+            },
             "$ai_http_status": 200,
             "$ai_input_tokens": usage_stats.get("prompt_tokens", 0),
             "$ai_output_tokens": usage_stats.get("completion_tokens", 0),
diff --git a/posthog/ai/providers/openai/openai_async.py b/posthog/ai/providers/openai/openai_async.py
index 93b8db2..ddce1ac 100644
--- a/posthog/ai/providers/openai/openai_async.py
+++ b/posthog/ai/providers/openai/openai_async.py
@@ -7,7 +7,7 @@
     raise ModuleNotFoundError("Please install OpenAI to use this feature: 'pip install openai'")
 
 from posthog.client import Client as PostHogClient
-from posthog.ai.utils import track_usage_async, get_model_params
+from posthog.ai.utils import call_llm_and_track_usage_async, get_model_params
 
 
 class AsyncOpenAI:
@@ -29,6 +29,15 @@ def __init__(
         self._openai_client = openai.AsyncOpenAI(**openai_config)
         self._posthog_client = posthog_client
 
+    def __getattr__(self, name: str) -> Any:
+        """
+        Expose all attributes of the underlying openai.AsyncOpenAI instance except for the 'chat' property,
+        which is replaced with a custom AsyncChatNamespace for usage tracking.
+        """
+        if name == "chat":
+            return self.chat
+        return getattr(self._openai_client, name)
+
     @property
     def chat(self) -> "AsyncChatNamespace":
         return AsyncChatNamespace(self._posthog_client, self._openai_client)
@@ -71,11 +80,10 @@ async def create(
         async def call_async_method(**call_kwargs):
             return await self._openai_client.chat.completions.create(**call_kwargs)
 
-        response = await track_usage_async(
+        response = await call_llm_and_track_usage_async(
             distinct_id, self._ph_client, posthog_trace_id, posthog_properties, call_async_method, **kwargs
         )
         return response
-    
 
     async def _create_streaming(
         self,
@@ -106,7 +114,9 @@ async def async_generator():
                 end_time = time.time()
                 latency = end_time - start_time
                 output = "".join(accumulated_content)
-                self._capture_streaming_event(distinct_id, posthog_trace_id, posthog_properties, kwargs, usage_stats, latency, output)
+                self._capture_streaming_event(
+                    distinct_id, posthog_trace_id, posthog_properties, kwargs, usage_stats, latency, output
+                )
 
         return async_generator()
 
@@ -120,7 +130,7 @@ def _capture_streaming_event(
         latency: float,
         output: str,
     ):
-        
+
         event_properties = {
             "$ai_provider": "openai",
             "$ai_model": kwargs.get("model"),
@@ -133,7 +143,7 @@ def _capture_streaming_event(
                         "role": "assistant",
                     }
                 ]
-            }, 
+            },
             "$ai_http_status": 200,
             "$ai_input_tokens": usage_stats.get("prompt_tokens", 0),
             "$ai_output_tokens": usage_stats.get("completion_tokens", 0),
@@ -148,4 +158,3 @@ def _capture_streaming_event(
                 event="$ai_generation",
                 properties=event_properties,
             )
-
diff --git a/posthog/ai/utils.py b/posthog/ai/utils.py
index d28629d..5be6483 100644
--- a/posthog/ai/utils.py
+++ b/posthog/ai/utils.py
@@ -1,4 +1,4 @@
-from typing import Any, Dict, AsyncGenerator, Callable, Optional
+from typing import Any, Dict, Callable, Optional
 import time
 from posthog.client import Client as PostHogClient
 
@@ -39,7 +39,7 @@ def format_response(response):
     return output
 
 
-def track_usage(
+def call_llm_and_track_usage(
     distinct_id: str,
     ph_client: PostHogClient,
     posthog_trace_id: Optional[str],
@@ -99,7 +99,7 @@ def track_usage(
     return response
 
 
-async def track_usage_async(
+async def call_llm_and_track_usage_async(
     distinct_id: str,
     ph_client: PostHogClient,
     posthog_trace_id: Optional[str],