---
title: "Find your generation in Conversations | Grafana Labs"
description: "Open Conversations in Agent Observability, set the time range to cover your run, and inspect the generation you captured in the Flow view."
---

> For a curated documentation index, see [llms.txt](/llms.txt). For the complete documentation index, see [llms-full.txt](/llms-full.txt).

# Find your generation in Conversations

Back in Grafana Cloud, the generation you captured in Milestone 5 should appear in **Conversations**. Open the list, confirm the time range covers your run, and open the conversation to inspect the generation.

Open Conversations and make the time range cover your run.

1. Go to **Agent Observability** > **Conversations**.
2. Check the time picker. Pick a range that covers the moment you ran the sample.
   
   For example, **Last 15 minutes**.
3. Click the refresh button to reload the list.
   
   The list shows a row for conversation `conv-1` with the agent, model (`openai/gpt-4o`), token count, cost, and duration from the generation you captured.
4. Click the `conv-1` row to open the detail view. **Flow** opens by default and shows the generation’s assistant output (`Hello` for the Python, JavaScript, Go, and Java samples), its timing data, and its token usage breakdown.

You found your generation in **Conversations**.

Nothing in the list? Check these in order:

1. The time range covers when you ran the sample. Widen it and refresh.
2. The sample completed and called the shutdown method last, so the queued generation was exported.
3. The five `AGENTO11Y_` variables were set in the same shell that ran the sample, and `AGENTO11Y_AUTH_TENANT_ID` is the **Instance ID** from Milestone 2, not the OTLP instance ID.
4. The access policy from Milestone 3 has the `sigil:write` scope and the token you exported belongs to that policy.