diff --git a/pkg/epp/framework/plugins/scheduling/scorer/prefix/README.md b/pkg/epp/framework/plugins/scheduling/scorer/prefix/README.md
index 8e8cfb16c2..df2ccf9aaf 100644
--- a/pkg/epp/framework/plugins/scheduling/scorer/prefix/README.md
+++ b/pkg/epp/framework/plugins/scheduling/scorer/prefix/README.md
@@ -31,6 +31,23 @@ The attribute is typically produced by the approximate prefix cache data produce
 
 This plugin does not define any plugin-specific parameters.
 
+### Agentic Workload Example
+
+For agentic workloads, sessions typically consist of long-running, multi-turn interactions where the system prompt is shared but session-specific context diverges rapidly.
+
+To optimize routing for these workloads, configure both the `prefix-cache-scorer` (with length awareness enabled) and the `session-affinity-scorer`. Length awareness allows shorter sessions to migrate freely when pods are overloaded, while session affinity keeps the rapidly diverging chat history sticky to the original pod:
+
+```yaml
+- type: prefix-cache-scorer
+  parameters:
+    matchLengthWeight: 0.8
+    matchLengthScaleTokens: 16000
+
+- type: session-affinity-scorer
+  parameters:
+    headerName: x-session-token
+```
+
 ## Operational notes
 
 - The scorer itself does not hash prompts or maintain cache state.
diff --git a/pkg/epp/framework/plugins/scheduling/scorer/sessionaffinity/README.md b/pkg/epp/framework/plugins/scheduling/scorer/sessionaffinity/README.md
index df67edc6c0..3ca3ed1c18 100644
--- a/pkg/epp/framework/plugins/scheduling/scorer/sessionaffinity/README.md
+++ b/pkg/epp/framework/plugins/scheduling/scorer/sessionaffinity/README.md
@@ -42,6 +42,21 @@ To support session affinity with PD disaggregation, configure two separate insta
 
 The decode instance uses the default behavior (writing the decode pod to `x-session-token`). The prefill instance uses `profileName: prefill` to look up the prefill pod from the scheduling results and write it to `x-session-token-prefill`. This ensures that subsequent requests in the same session target both the same prefill pod and the same decode pod.
 
+### Agentic Workload Example
+
+For agentic workloads (long-running, multi-turn interactions), it is highly recommended to pair this scorer with the `prefix-cache-scorer` (with length awareness enabled) to ensure rapidly diverging chat history remains sticky to the original pod, while allowing shorter new sessions to migrate to less loaded pods:
+
+```yaml
+- type: prefix-cache-scorer
+  parameters:
+    matchLengthWeight: 0.8
+    matchLengthScaleTokens: 16000
+
+- type: session-affinity-scorer
+  parameters:
+    headerName: x-session-token
+```
+
 ## Relationship to the session affinity filter
 
 The [session affinity filter](../../filter/sessionaffinity/README.md) (`session-affinity-filter`) provides the same affinity behavior as a hard constraint and writes the same response header. Configuring both alongside the scorer is unnecessary and can be misleading; see [Relationship to the session affinity scorer](../../filter/sessionaffinity/README.md#relationship-to-the-session-affinity-scorer) for details. Use the scorer for a soft preference that can be outweighed by other scorers, or the filter for a hard pin.