diff --git a/projects/rhaiis/orchestration/ci.py b/projects/rhaiis/orchestration/ci.py
index 0e76806f0..3508cdefd 100755
--- a/projects/rhaiis/orchestration/ci.py
+++ b/projects/rhaiis/orchestration/ci.py
@@ -9,9 +9,11 @@
 import prepare_rhaiis
 import test_rhaiis
 
+from projects.core.agentic.config_review import trigger_config_review_for_ci
+from projects.core.agentic.on_failure import agent_review_on_failure
 from projects.core.ci_entrypoint.fournos_resolve import create_fournos_resolve_entrypoint
 from projects.core.library import ci as ci_lib
-from projects.core.library import vault
+from projects.core.library import env, vault
 from projects.core.library.export import caliper_export_entrypoint
 from projects.rhaiis.orchestration import runtime_config
 
@@ -139,8 +141,10 @@ def prepare(ctx):
 @main.command()
 @click.pass_context
 @ci_lib.safe_ci_entrypoint
+@agent_review_on_failure
 def test(ctx):
     """Test phase - Deploy model, run benchmarks, capture results."""
+    trigger_config_review_for_ci(env.BASE_ARTIFACT_DIR, async_mode=True)
     return test_rhaiis.test()
 
 
diff --git a/projects/rhaiis/orchestration/config.d/models.yaml b/projects/rhaiis/orchestration/config.d/models.yaml
index 59739985c..ef50d186b 100644
--- a/projects/rhaiis/orchestration/config.d/models.yaml
+++ b/projects/rhaiis/orchestration/config.d/models.yaml
@@ -5,6 +5,13 @@ qwen3-0_6b:
   vllm_args:
     tensor-parallel-size: 1
 
+# Dashboard override stub (set hf_model_id and vllm_args at runtime)
+custom:
+  name: custom
+  hf_model_id: ""
+  vllm_args:
+    tensor-parallel-size: 1
+
 # Llama-4 Scout (MoE)
 llama-4-scout:
   name: Llama-4-Scout-17B-16E-Instruct
diff --git a/projects/rhaiis/orchestration/config.d/workloads.yaml b/projects/rhaiis/orchestration/config.d/workloads.yaml
index c03e32bc2..b1c277004 100644
--- a/projects/rhaiis/orchestration/config.d/workloads.yaml
+++ b/projects/rhaiis/orchestration/config.d/workloads.yaml
@@ -7,25 +7,21 @@ profile1:
   data: "prompt_tokens=1000,output_tokens=1000"
   rates: [1,50,100,200,300]
   max_seconds: 450
-  rampup: 10
 
 profile2:
   data: "prompt_tokens=512,prompt_tokens_stdev=128,prompt_tokens_min=1,prompt_tokens_max=1024,output_tokens=2048,output_tokens_stdev=512,output_tokens_min=1,output_tokens_max=4096"
   rates: [1,50,100,200,300]
   max_seconds: 450
-  rampup: 10
 
 profile3:
   data: "prompt_tokens=2048,output_tokens=128"
   rates: [1, 50, 100, 200, 300]
   max_seconds: 450
-  rampup: 10
 
 profile4:
   data: "prompt_tokens=8000,output_tokens=1000"
   rates: [1,25,50,75,100]
   max_seconds: 450
-  rampup: 10
   samples: 50
 
 profile5:
@@ -44,3 +40,8 @@ profile7:
   rates: [1, 50, 100, 200, 300]
   max_seconds: 450
 
+# Dashboard override stub (set data, rates, max_seconds at runtime)
+custom:
+  data: "prompt_tokens=1000,output_tokens=1000"
+  rates: [1]
+  max_seconds: 450
diff --git a/projects/rhaiis/orchestration/config.yaml b/projects/rhaiis/orchestration/config.yaml
index dd7206936..320a42cf3 100644
--- a/projects/rhaiis/orchestration/config.yaml
+++ b/projects/rhaiis/orchestration/config.yaml
@@ -3,6 +3,7 @@ vaults:
 - psap-forge-mlflow-export
 - psap-forge-aws-s3-export
 - psap-forge-dashboard-s3
+- psap-models-corp-rh
 
 benchmarks:
   guidellm:
@@ -77,3 +78,11 @@ caliper:
           tags: {}
           parameters: {}
           metrics: {}
+
+agentic:
+  enabled: true
+  model_key: qwen-3-6-35b
+  on_failure:
+    enabled: true
+  config_review:
+    enabled: true