diff --git a/ai_client.py b/ai_client.py
index 428a595..89c2db0 100644
--- a/ai_client.py
+++ b/ai_client.py
@@ -1764,6 +1764,43 @@ def _send_deepseek(md_content: str, user_message: str, base_dir: str,
         raise _classify_deepseek_error(e) from e
 
 
+def run_tier4_analysis(stderr: str) -> str:
+    """
+    Stateless Tier 4 QA analysis of an error message.
+    Uses gemini-2.5-flash-lite to summarize the error and suggest a fix.
+    """
+    if not stderr or not stderr.strip():
+        return ""
+
+    try:
+        _ensure_gemini_client()
+        prompt = (
+            f"You are a Tier 4 QA Agent specializing in error analysis.\n"
+            f"Analyze the following stderr output from a PowerShell command:\n\n"
+            f"```\n{stderr}\n```\n\n"
+            f"Provide a concise summary of the failure and suggest a fix in approximately 20 words."
+        )
+
+        # Use flash-lite for cost-effective stateless analysis
+        model_name = "gemini-2.5-flash-lite"
+
+        # We don't use the chat session here to keep it stateless
+        resp = _gemini_client.models.generate_content(
+            model=model_name,
+            contents=prompt,
+            config=types.GenerateContentConfig(
+                temperature=0.0,
+                max_output_tokens=150,
+            )
+        )
+
+        analysis = resp.text.strip()
+        return analysis
+    except Exception as e:
+        # We don't want to crash the main loop if QA analysis fails
+        return f"[QA ANALYSIS FAILED] {e}"
+
+
 # ------------------------------------------------------------------ unified send
 
 def send(
diff --git a/tests/test_tier4_interceptor.py b/tests/test_tier4_interceptor.py
index 10e6453..d03b61d 100644
--- a/tests/test_tier4_interceptor.py
+++ b/tests/test_tier4_interceptor.py
@@ -100,3 +100,29 @@ def test_run_powershell_optional_qa_callback():
         
         assert "STDERR:\nerror" in output
         assert "EXIT CODE: 1" in output
+
+def test_end_to_end_tier4_integration():
+    """
+    Verifies that shell_runner.run_powershell correctly uses ai_client.run_tier4_analysis.
+    """
+    import ai_client
+    
+    script = "Invoke-Item non_existent_file"
+    base_dir = "."
+    stderr_content = "Invoke-Item : Cannot find path 'C:\\non_existent_file' because it does not exist."
+    
+    mock_result = MagicMock()
+    mock_result.stdout = ""
+    mock_result.stderr = stderr_content
+    mock_result.returncode = 1
+    
+    expected_analysis = "Path does not exist. Verify the file path and ensure the file is present before invoking."
+    
+    with patch("subprocess.run", return_value=mock_result), \
+         patch("shutil.which", return_value="powershell.exe"), \
+         patch("ai_client.run_tier4_analysis", return_value=expected_analysis) as mock_analysis:
+        
+        output = run_powershell(script, base_dir, qa_callback=ai_client.run_tier4_analysis)
+        
+        mock_analysis.assert_called_once_with(stderr_content)
+        assert f"QA ANALYSIS:\n{expected_analysis}" in output