diff --git a/docs/error-tracking.md b/docs/error-tracking.md
index 41349bccba..1bd417d768 100644
--- a/docs/error-tracking.md
+++ b/docs/error-tracking.md
@@ -31,12 +31,27 @@ Privacy constraints for agent monitoring:
 
 - `recordInputs` / `recordOutputs` are **false** on the wrap and `dataCollection.genAI` is `{ inputs: false, outputs: false }` in both runtime configs — prompts, clinical queries, source evidence, generated answers, and embedding inputs are never recorded.
 - `privacySafeTransactionEvent` allowlists only gen_ai metadata attributes (system, operation name, request/response model, response id, finish reasons, token usage, conversation id) and rebuilds gen_ai span descriptions as `<operation> <model>` from those attributes. Message, prompt, tool-payload, and embedding-input attributes are stripped on export even if a future SDK version records them.
-- Each answer request (`/api/answer` and `/api/answer/stream`, summaries included) calls `Sentry.setConversationId(<interactionId>)` — the request's synthetic UUID — so the embedding/generation calls of one request group into one conversation without carrying any query text.
+- Each answer / stream / document-summarize request calls `Sentry.setConversationId(<interactionId>)` **before** OpenAI work — the request's synthetic UUID — so the embedding/generation calls of one request group into one conversation without carrying any query text.
 - User identification (`Sentry.setUser`) is deliberately **not** wired: the committed privacy boundary strips `user` from every outgoing event (see the tests), and linking clinical-query telemetry to an identity would need its own governance review first.
 - `responses.parse` (schema-parsed generation) is not in the SDK's instrumentation registry and emits no gen_ai span; `responses.create` and `embeddings.create` are covered.
 
 Rollback matches tracing: set `SENTRY_TRACES_SAMPLE_RATE=0` (agent spans stop; error capture stays) or remove `SENTRY_DSN`. Raising the sample rate above the 0.1 default captures a larger share of answer requests in the agents view and is an operator decision.
 
+### Sentry “Agent Monitoring” wizard mapping
+
+The product wizard’s copy is generic. Map it to this repo as follows — do **not** paste the wizard’s sample code:
+
+| Wizard step                      | This repo                                                                                                                                                                |
+| -------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| Install `@sentry/nextjs` ≥ 10.67 | repository dependency `@sentry/nextjs@^10.69.0` (lockfile resolves `10.69.0`); `10.67` is the minimum supported version; do not switch to pnpm                           |
+| `Sentry.init` + tracing          | `src/sentry.server.config.ts` / `src/sentry.edge.config.ts`; DSN from `SENTRY_DSN` only (never hardcode); default `tracesSampleRate` 0.1 via `SENTRY_TRACES_SAMPLE_RATE` |
+| `dataCollection.genAI`           | Explicitly `{ inputs: false, outputs: false }`                                                                                                                           |
+| `instrumentOpenAiClient`         | `instrumentOpenAIClientForAgentMonitoring` in `createOpenAIClient()` — `recordInputs`/`recordOutputs` **false**                                                          |
+| `setConversationId`              | `setAgentConversationId(interactionId)` on `/api/answer`, `/api/answer/stream`, and `/api/documents/[id]/summarize`                                                      |
+| `setUser` (optional)             | **Rejected** — scrubbers strip `user`; do not wire                                                                                                                       |
+
+Leaving wizard checkboxes for “record inputs/outputs” or “identify users” unchecked is expected and correct for this clinical app.
+
 ### Structured logs (Sentry Logs)
 
 Server/edge `enableLogs` turns on when `SENTRY_DSN` is set (disable with `SENTRY_ENABLE_LOGS=false`). There is still no browser logging path and no `consoleLoggingIntegration` — console output is too easy to leak clinical text.
diff --git a/src/app/api/documents/[id]/summarize/route.ts b/src/app/api/documents/[id]/summarize/route.ts
index e2e0fc7564..25e3f90434 100644
--- a/src/app/api/documents/[id]/summarize/route.ts
+++ b/src/app/api/documents/[id]/summarize/route.ts
@@ -10,6 +10,7 @@ import { logAnswerDiagnostics } from "@/lib/answer-telemetry";
 import { answerFeedbackMetadata } from "@/lib/answer-feedback-token";
 import { jsonError } from "@/lib/http";
 import { consumeApiRateLimit, rateLimitJsonResponse } from "@/lib/api-rate-limit";
+import { setAgentConversationId } from "@/lib/observability/agent-monitoring";
 import { createAdminClient } from "@/lib/supabase/admin";
 import { AuthenticationError, requireAuthenticatedUser, unauthorizedResponse } from "@/lib/supabase/auth";
 import { parseRouteParams } from "@/lib/validation/params";
@@ -39,6 +40,10 @@ export async function POST(request: Request, { params }: { params: Promise<{ id:
     const rateLimit = await consumeApiRateLimit({ supabase, ownerId: user.id, bucket: "document_summarize" });
     if (rateLimit.limited)
       return rateLimitJsonResponse("Too many document summary requests. Retry shortly.", rateLimit);
+    // Group this request's LLM calls into one Sentry agent-monitoring conversation
+    // before any OpenAI work starts. Synthetic UUID only — never document/query text.
+    const interactionId = randomUUID();
+    setAgentConversationId(interactionId);
     const answer = await summarizeDocument(id, user.id, { signal: request.signal });
     const governedResponse = buildGovernedAnswerClientResponse(answer);
     logAnswerDiagnostics({
@@ -47,7 +52,6 @@ export async function POST(request: Request, { params }: { params: Promise<{ id:
       ownerId: user.id,
       answer: governedResponse.telemetryAnswer,
     });
-    const interactionId = randomUUID();
     return NextResponse.json({
       ...governedResponse.payload,
       ...answerFeedbackMetadata(interactionId, governedResponse.payload.answer),
diff --git a/tests/summarize-agent-conversation.test.ts b/tests/summarize-agent-conversation.test.ts
new file mode 100644
index 0000000000..98ff029716
--- /dev/null
+++ b/tests/summarize-agent-conversation.test.ts
@@ -0,0 +1,95 @@
+import { afterEach, describe, expect, it, vi } from "vitest";
+
+const setAgentConversationId = vi.fn();
+const summarizeDocument = vi.fn(async () => ({
+  answer: "Summary body.",
+  grounded: true,
+  confidence: "high",
+  citations: [],
+  smartPanel: { query: "summary" },
+  smartApiPlan: { displayMode: "direct" },
+}));
+
+vi.mock("@/lib/observability/agent-monitoring", () => ({
+  setAgentConversationId,
+}));
+
+vi.mock("@/lib/demo-data", () => ({
+  isDemoMode: () => false,
+  getDemoDocument: () => null,
+  demoSummary: () => ({}),
+}));
+
+vi.mock("@/lib/env", () => ({
+  isDemoMode: () => false,
+}));
+
+vi.mock("@/lib/rag/rag", () => ({
+  summarizeDocument,
+}));
+
+vi.mock("@/lib/answer-response", () => ({
+  buildGovernedAnswerClientResponse: (answer: { answer: string }) => ({
+    payload: { answer: answer.answer, grounded: true },
+    telemetryAnswer: answer,
+  }),
+  buildGovernedDemoAnswerClientResponse: (payload: unknown) => payload,
+}));
+
+vi.mock("@/lib/answer-telemetry", () => ({
+  logAnswerDiagnostics: vi.fn(),
+}));
+
+vi.mock("@/lib/answer-feedback-token", () => ({
+  answerFeedbackMetadata: (interactionId: string) => ({ interactionId }),
+}));
+
+vi.mock("@/lib/api-rate-limit", () => ({
+  consumeApiRateLimit: vi.fn(async () => ({
+    limited: false,
+    limit: 12,
+    remaining: 11,
+    retryAfterSeconds: 60,
+    resetAt: new Date().toISOString(),
+  })),
+  rateLimitJsonResponse: vi.fn(),
+}));
+
+vi.mock("@/lib/supabase/admin", () => ({
+  createAdminClient: () => ({}),
+}));
+
+vi.mock("@/lib/supabase/auth", () => ({
+  AuthenticationError: class AuthenticationError extends Error {},
+  requireAuthenticatedUser: vi.fn(async () => ({ id: "11111111-1111-4111-8111-111111111111" })),
+  unauthorizedResponse: () => new Response(null, { status: 401 }),
+}));
+
+describe("document summarize Sentry conversation id", () => {
+  afterEach(() => {
+    setAgentConversationId.mockClear();
+    summarizeDocument.mockClear();
+    vi.resetModules();
+  });
+
+  it("sets the agent conversation id before any LLM summarize work", async () => {
+    const { POST } = await import("../src/app/api/documents/[id]/summarize/route");
+    const documentId = "aaaaaaaa-aaaa-4aaa-8aaa-aaaaaaaaaaaa";
+
+    const response = await POST(
+      new Request(`http://localhost/api/documents/${documentId}/summarize`, { method: "POST" }),
+      {
+        params: Promise.resolve({ id: documentId }),
+      },
+    );
+
+    expect(response.status).toBe(200);
+    expect(setAgentConversationId).toHaveBeenCalledTimes(1);
+    expect(setAgentConversationId.mock.invocationCallOrder[0]).toBeLessThan(
+      summarizeDocument.mock.invocationCallOrder[0]!,
+    );
+    const body = (await response.json()) as { interactionId?: string };
+    expect(body.interactionId).toMatch(/^[0-9a-f]{8}-[0-9a-f]{4}-[1-8][0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i);
+    expect(setAgentConversationId).toHaveBeenCalledWith(body.interactionId);
+  });
+});