[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC)：2025-08-18。"],[[["\u003cp\u003eThe conversation history tool allows users to browse and analyze real conversations between their agent and end-users to evaluate agent performance and debug issues.\u003c/p\u003e\n"],["\u003cp\u003eFlow and page analysis provides insights into traffic patterns and escalations within the agent's flow graph, helping identify high-traffic paths and problem areas.\u003c/p\u003e\n"],["\u003cp\u003eUsers can filter conversations and flow analysis data for maximum flexibility in studying a specific subset of agent traffic by intent, user/agent utterances, flags, and more.\u003c/p\u003e\n"],["\u003cp\u003eThe tool has limitations, including slow loading times for large agents, which can be improved by using intent filters and shorter date ranges, as well as a 500 turn cap per conversation.\u003c/p\u003e\n"],["\u003cp\u003eConversation history can be exported to a CSV file (limited to 50 conversations) or accessed via BigQuery for larger batches, with API access also available for programmatic use.\u003c/p\u003e\n"]]],[],null,["# Conversation history\n\nThe conversation history tool provides an interface\nfor browsing and analyzing actual production conversations\nbetween your agent and end-users.\nThis can be useful to evaluate the functional performance of your agent\nor to debug issues.\n\nFlow and page analysis helps you understand\ntraffic and escalations through the flow graph of an agent. This enables\nagent builders to analyze traffic flow (high and low traffic paths), discover\nissue hotspots (for example, to pinpoint high escalations), and drill down to\nthe specific pages and intents causing issues. Flow and page analysis can be\nused along with conversation history filters, to allow for maximum flexibility\nin studying a relevant subset of traffic.\n\nLimitations\n-----------\n\nLoading the Flow Analysis Table and Flow Analysis Graph views\ncan be slow for large agents.\nTo improve load times,\nuse an intent filter and a one week date range.\n\nAt most 500 turns will be logged for each conversation.\n\nPermissions required\n--------------------\n\nUsing this tool requires the **dialogflow.conversations** permission.\nThis permission can be added to a\n[custom role](/iam/docs/creating-custom-roles).\nIt is also available in the\n[Admin role](/dialogflow/cx/docs/concept/access-control#df-console),\nas well as\n[other Conversational Agents (Dialogflow CX) roles](/iam/docs/understanding-roles#dialogflow-roles).\n\nEnable and disable conversation history\n---------------------------------------\n\nThe conversation history tool uses data from interaction logs.\nYou can enable and disable conversation history with the\n[Enable interaction logging](/dialogflow/cx/docs/concept/agent-settings#settings-general)\nagent setting.\n\nYou can also control how long data is retained by\nusing the `retention_window_days`\n[security setting](/dialogflow/cx/docs/concept/security-settings).\n\nBrowsing the conversation history\n---------------------------------\n\nTo use the conversation history tool:\n\n1. Open the [Dialogflow CX console](https://dialogflow.cloud.google.com/cx/projects).\n2. Choose your project.\n3. Select your agent.\n4. Select the **Manage** tab.\n5. Click **Conversation History**.\n6. Conversations are listed in a table.\n7. Optionally apply filters using the **filter** control above the table.\n8. Select one of the tabs, each of which are described later in this document.\n\nDownload conversation history to a CSV file\n-------------------------------------------\n\n| **Note:** The maximum number of conversations that can be downloaded to a CSV file is **50** . You can apply filters in the **Filter conversations** field at the top of the **Conversation History** page to select only the specific conversations you want to export. To export larger batches of conversations, you must [use BigQuery](/dialogflow/cx/docs/concept/export-bq).\n\n1. Open the [Dialogflow CX console](https://dialogflow.cloud.google.com/cx/projects).\n2. Choose your project.\n3. Select your agent.\n4. Select the **Manage** tab.\n5. Click **Conversation History**.\n6. Conversation sessions are listed in a table.\n7. Select a conversation session, then click the **open conversation in\n preview panel** button:\n\n8. Click **Export all conversations** . A CSV file containing the conversations\n will automatically download to your computer. **The maximum number of\n conversations that can be downloaded is 50**. If the conversation session\n you selected contains more than 50 conversations, the download will fail.\n\nConversations\n-------------\n\nConversations are shown in two views:\nconversation summary view and conversation details view.\n\n### Conversation summary view\n\nThe conversation summary view lists all conversations.\nThe following metadata is provided for each conversation:\n\nYou can filter the results by metadata (except environment and duration)\nand the following additional filter options:\n\n### Conversation details view\n\nYou can browse the turn-by-turn view of a specific conversation\nin the conversation details view.\nEach conversational turn provides the end-user message,\nthe agent message,\nand the following metadata:\n\n### Example use case\n\nThe tool can be used to identify agent issues that result\nin escalations to a human agent.\nFrom the conversation summary view,\nyou can apply the **Is Live Agent Handoff** filter for **Yes**.\nExamine some of these conversations\nand try to find common patterns.\nFor example,\nyou may find that in most of the conversations,\nthere is a no-match flag for turns on a specific page.\nThis would indicate that an intent route needs to be created,\nor that an existing intent route is not in scope when it needs to be.\n\nFlow analysis table\n-------------------\n\nThe **Flow Analysis - Table** tab\nis used to analyze traffic and exits\nin flows and pages in tabular form.\n\nThe following metrics are shown:\n\nClicking the\nlist list_alt button\nopens a sample of relevant conversations for the row.\nYou can read through the samples to better understand specific and nuanced issues.\nFor example,\nyou can examine why escalations are happening.\nTo view the full details of a sample,\nclick the **Conversation details** link.\n\nClicking **Detailed Stats** for a flow\nopens the traffic stats for flows.\nThis shows flow transition stats from and to the selected flow.\n\nThe following metrics are shown:\n\nClicking any flow row drills down into the pages that compromise that flow.\n\nThe following metrics are shown:\n\nClicking **Detailed Stats** for a page\nopens the traffic stats for pages.\nThis shows page transition stats from and to the selected page.\n\nThe following metrics are shown:\n\nFlow analysis graph\n-------------------\n\nThe **Flow Analysis - Graph** tab\nis used to analyze traffic and exits\nin flows and pages in graph form.\n\nClicking on any flow drills down into the traffic for\nthe pages within the flow.\n\nYou can use the zoom slider to vary the level of detail that you see on the\ngraph. Low volume paths are dropped from the visualization at lower settings.\nEven at higher settings, very low volume paths are always hidden.\n\nMetrics are color-coded to easily find high severity issues.\n\nClicking the\nlist list_alt button\nshows relevant conversations.\n\nClicking on any flow takes you to the pages within that flow. Here you get more\noverlay metrics:\nrelative traffic, no match rate, escalation rate, and exit rate.\n\nAccess conversation history with the API\n----------------------------------------\n\n|\n| **Preview**\n|\n|\n| This product or feature is subject to the \"Pre-GA Offerings Terms\" in the General Service Terms section\n| of the [Service Specific Terms](/terms/service-terms#1).\n|\n| Pre-GA products and features are available \"as is\" and might have limited support.\n|\n| For more information, see the\n| [launch stage descriptions](/products#product-launch-stages).\n\nYou can access conversation history with the API.\nSee the\n[V3beta1 reference documentation](/dialogflow/cx/docs/reference/rpc/google.cloud.dialogflow.cx.v3beta1#conversationhistory).\n\nPrivacy\n-------\n\nWhen you enable conversation history, Google collects and stores your\nconversation data for a period of time before permanently deleting it. The\ndefault is 365 days, but you can shorten this time period using the\n`retention_window_days` field in\n[`SecuritySettings`](https://cloud.google.com/dialogflow/cx/docs/reference/rpc/google.cloud.dialogflow.cx.v3beta1#securitysettings). \n\nYou retain full ownership of your data while it's stored. Google won't use\nyour data or access it for any purposes other than customer support."]]
