diff --git a/mlflow/models/display_utils.py b/mlflow/models/display_utils.py
index 37726fecf22fb..264be17195707 100644
--- a/mlflow/models/display_utils.py
+++ b/mlflow/models/display_utils.py
@@ -6,7 +6,17 @@
 from mlflow.utils import databricks_utils
 
 
+def _is_input_string(inputs: schema.Schema) -> bool:
+    return (
+        not inputs.has_input_names()
+        and len(inputs.input_types()) == 1
+        and inputs.input_types()[0] == schema.DataType.string
+    )
+
+
 def _is_input_agent_compatible(inputs: schema.Schema) -> bool:
+    if _is_input_string(inputs):
+        return True
     if not inputs.has_input_names():
         return False
     messages = inputs.input_dict().get("messages")
diff --git a/tests/models/test_display_utils.py b/tests/models/test_display_utils.py
index dbea5b49913c9..c77ff80c03f93 100644
--- a/tests/models/test_display_utils.py
+++ b/tests/models/test_display_utils.py
@@ -62,6 +62,11 @@ def test_should_render_eval_template_with_vanilla_string(enable_databricks_env):
     assert should_render_agent_eval_template(signature)
 
 
+def test_should_render_eval_template_with_string_input(enable_databricks_env):
+    signature = infer_signature("A vanilla string input", _STRING_RESPONSE)
+    assert should_render_agent_eval_template(signature)
+
+
 def test_should_not_render_eval_template_generic_signature(enable_databricks_env):
     signature = infer_signature({"input": "string"}, {"output": "string"})
     assert not should_render_agent_eval_template(signature)