| \n\nApp\n\n | \n\nApplication using NLU. For example: **VA** \\(Virtual Agent\\).\n\n |
| \n\nApp Document\n\n | \n\nsys\\_id of the application record. For example: **sys\\_cs\\_topic.sys\\_id**.\n\n |
| \n\nApp document table\n\n | \n\nTable of the application record. For example: **sys\\_cs\\_topic** \\(Topics table\\).\n\n |
| \n\nAuto-Selected\n\n | \n\nBoolean indicating whether the system detected only one intent above the prediction threshold, and therefore the application chose to run the topic that matched the same intent: **true** or **false**.\n\n |
| \n\nCurrent Intent\n\n | \n\nNLU intent that was used by the application when the NLU prediction was submitted. This field can be empty \\(blank\\). For example, an intent used in the Virtual Agent application could be **Change password**.\n\n |
| \n\nDomain\n\n | \n\nThe intent's domain. For example, Global or ITSM.\n\n |
| \n\nExternal Model Display Name\n\n | \n\nName of the external NLU model. For example, the name of the IBM Watson Assistant NLU application.\n\n |
| \n\nIntent Switch Causer\n\n | \n\nName of the originating topic that prompts the user for confirmation. There are two possible scenarios for values in this column:- The user is prompted for confirmation on the initial intent match for their utterance \\(**Ask user if topic VA chose is correct** toggle switch is enabled in [NLU Settings](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/conversational-interfaces/virtual-agent/configure-nlu-settings.md)\\).\n\nIf the user is prompted for an initial utterance, this value may be the Greetings topic. If the value is **empty**, the user either was not prompted for the initial match, or the user rejected the initially matched topic intent.\n\n- The user is prompted for confirmation for an intent switch in mid-topic \\(**Ask user if switched topic is correct** toggle switch is enabled in [NLU Settings](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/conversational-interfaces/virtual-agent/configure-nlu-settings.md)\\).\n\nIf the user is prompted for a switch in mid-topic, this value is the name of the topic they were viewing at the time of the switch.\n\n **Note:** System tasks may also display here, which you can ignore. For example: **\\_\\_silent\\_TerminateGoal\\_a3817a5f7ca7439b8**.\n\n |
| \n\nIntent Switch Causer current Task\n\n | \n\nName of the task \\(node\\) that the user switches from.**Note:** System tasks may also display here, which you can ignore. For example: **\\_\\_silent\\_TerminateGoal\\_a3817a5f7ca7439b8**.\n\n |
| \n\nMapped\n\n | \n\nBoolean indicating whether the application \\(Virtual Agent\\) considered the prediction to be usable by matching the intent to an available topic: **true** or **false**.\n\n |
| \n\nModel Display Name\n\n | \n\nName of the ServiceNow NLU model.\n\n |
| \n\nModel ID\n\n | \n\nID of the NLU model that the prediction was submitted for.\n\n |
| \n\nModel Name\n\n | \n\nName of the NLU model that the prediction was submitted for.\n\n |
| \n\nNLU Topic Discovery Outcome\n\n | \n\nPrediction result for initial utterance in a conversation as part of topic discovery:- **AI Search Fallback**: No prediction was made. AI Search was used as the fallback.\n- **Correct**: A single prediction was made. The user confirmed the topic was correct.\n- **Correct Among Multiple**: Multiple predictions were made. This result indicates that the user selected this prediction. Other prediction records are marked as Uncategorized.\n- **Exited with no confirmation**: The user exits with no confirmation.\n- **Incorrect**: A single prediction was made. The user confirmed the topic was incorrect.\n- **Incorrect Among Multiple**: Multiple predictions were made, but the user did not select any of the suggested topics. This result indicates that this prediction had the highest Prediction Confidence value. The other prediction records are marked as Uncategorized.\n- **Setup Topic Prediction**: Setup topic related prediction was made without using AI Search.\n- **Skipped**: No prediction was made.\n- **Skipped with New Utterance**: When the Topic menu displayed, the user typed a new utterance.\n- **Uncategorized**: None of the above conditions were met, or the user abandoned the conversation.\n\n\n |
| \n\nNLU Model Language\n\n | \n\nLanguage of the NLU model used for the prediction.\n\n |
| \n\nPrediction\n\n | \n\nName of the predicted intent.\n\n |
| \n\nPrediction Audit Log\n\n | \n\nReference to the Open NLU Predict Logs \\[open\\_nlu\\_predict\\_log\\] table.\n\n |
| \n\nPrediction Confidence\n\n | \n\nConfidence value of the prediction.\n\n |
| \n\nRequester Session Language\n\n | \n\nLanguage code of the requester session language when Dynamic Translation is enabled.For example, if the user types in French, the code **fr** is stored. If language detection is enabled and the user enters text in a language that is not the default, Virtual Agent asks the user to confirm their language. If confirmed, the detected language becomes the default value for the session.\n\n |
| \n\nSelected\n\n | \n\nBoolean indicating whether the multiple predicted intents shown to end users as matched topics were actually selected by the end users: **true** or **false**.\n\n |
| \n\nShown\n\n | \n\nBoolean indicating whether the application \\(Virtual Agent\\) displayed the predicted intent as a choice for the end user to select: **true** or **false**.\n\n |
| \n\nTopic Switched\n\n | \n\nName of the topic, if the user was prompted for confirmation \\(regardless of their choice, yes or no\\).\n\n |
| \n\nUtterance\n\n | \n\nPhrase entered by the end user in the chat client window.\n\n |
\n