diff --git a/src/groundlight/client.py b/src/groundlight/client.py
index 65e0b97b..c34f65f1 100644
--- a/src/groundlight/client.py
+++ b/src/groundlight/client.py
@@ -43,6 +43,7 @@ class Groundlight:
     """
 
     DEFAULT_WAIT: float = 30.0
+    DEFAULT_PATIENCE: float = 60.0
 
     POLLING_INITIAL_DELAY = 0.25
     POLLING_EXPONENTIAL_BACKOFF = 1.3  # This still has the nice backoff property that the max number of requests
@@ -223,6 +224,57 @@ def submit_image_query(  # noqa: PLR0913 # pylint: disable=too-many-arguments
             image_query = self.wait_for_confident_result(image_query, confidence_threshold=threshold, timeout_sec=wait)
         return self._fixup_image_query(image_query)
 
+    def ask_confident(self):
+        pass
+
+    def ask_ml(self):
+        pass
+
+    def ask_async(
+        self,
+        detector: Union[Detector, str],
+        image: Union[str, bytes, Image.Image, BytesIO, BufferedReader, np.ndarray],
+        patience_time: float = DEFAULT_PATIENCE,
+        human_review: Optional[str] = None,
+    ) -> ImageQuery:
+        """Sends an image to Groundlight without waiting for a response.
+        :param detector: the Detector object, or string id of a detector like `det_12345`
+        :param image: The image, in several possible formats:
+          - filename (string) of a jpeg file
+          - byte array or BytesIO or BufferedReader with jpeg bytes
+          - numpy array with values 0-255 and dimensions (H,W,3) in BGR order
+            (Note OpenCV uses BGR not RGB. `img[:, :, ::-1]` will reverse the channels)
+          - PIL Image
+          Any binary format must be JPEG-encoded already.  Any pixel format will get
+          converted to JPEG at high quality before sending to service.
+        :param patience_time: How long Groundlight will work to answer the query.
+        :param human_review: If `None` or `DEFAULT`, send the image query for human review
+            only if the ML prediction is not confident.
+            If set to `ALWAYS`, always send the image query for human review.
+            If set to `NEVER`, never send the image query for human review.
+        """
+
+        detector_id = detector.id if isinstance(detector, Detector) else detector
+
+        image_bytesio: ByteStreamWrapper = parse_supported_image_types(image)
+
+        params = {"detector_id": detector_id, "body": image_bytesio}
+        if patience_time == 0:
+            params["patience_time"] = self.DEFAULT_PATIENCE
+        else:
+            params["patience_time"] = patience_time
+
+        if human_review is not None:
+            params["human_review"] = human_review
+
+        # If no inspection_id is provided, we submit the image query using image_queries_api (autogenerated via OpenAPI)
+        # However, our autogenerated code does not currently support inspection_id, so if an inspection_id was
+        # provided, we use the private API client instead.
+
+        raw_image_query = self.image_queries_api.submit_image_query(**params)
+        image_query = ImageQuery.parse_obj(raw_image_query.to_dict())
+        return self._fixup_image_query(image_query)
+
     def wait_for_confident_result(
         self,
         image_query: Union[ImageQuery, str],
diff --git a/src/groundlight/todo.txt b/src/groundlight/todo.txt
new file mode 100644
index 00000000..8aeeda58
--- /dev/null
+++ b/src/groundlight/todo.txt
@@ -0,0 +1,10 @@
+out of scope:
+    changing the behavior of submit_image_query - this is a breaking change that will require a lot of love and planning. I see submit_image_query as the heavy duty utility function, being a superset of the three primary functions we're introducing
+TODO:
+    write
+        aks_ml
+        aks confident
+    add test cases
+        ask_async
+        aks_ml
+        aks confident