The output schema and column definitions can be found in the table below.
| Name | Type | Definition |
|---|---|---|
| conversationName | STRING | Name of the conversation resource. |
| audioFileUri | STRING | Location of the audio file on Google Cloud Storage. |
| agentId | STRING | The user-provided identifier for the agent who handled the conversation. |
| startTimestampUtc | INTEGER | The time in seconds at which the conversation started. |
| loadTimestampUtc | INTEGER | The time in seconds at which the conversation was loaded into Insights. |
| year | INTEGER | Year date part of loadTimestampUtc. |
| month | INTEGER | Month date part of loadTimestampUtc. |
| day | INTEGER | Day date part of loadTimestampUtc. |
| durationNanos | INTEGER | Conversation duration in nanoseconds. |
| silenceNanos | INTEGER | Number of nanoseconds calculated to be in silence. |
| silencePercentage | FLOAT | Percentage of the total conversation spent in silence. |
| agentSpeakingPercentage | FLOAT | Percentage of the conversation with the agent speaking. |
| clientSpeakingPercentage | FLOAT | Percentage of the conversation with the client speaking. |
| agentSentimentScore | FLOAT | Agent sentiment score between -1.0 (negative) and 1.0 (positive). |
| agentSentimentMagnitude | FLOAT | A non-negative number from zero to infinity that represents the abolute magnitude of the agent sentiment regardless of score. |
| clientSentimentScore | FLOAT | Client sentiment score between -1.0 (negative) and 1.0 (positive). |
| clientSentimentMagnitude | FLOAT | A non-negative number from zero to infinity that represents the abolute magnitude of client sentiment regardless of score. |
| transcript | STRING | The complete text transcript of the conversation. |
| turnCount | INTEGER | The number of turns taken in the conversation. |
| languageCode | STRING | Language code of the conversation. |
| medium | STRING | Medium of the conversation. |
| issues | REPEATED RECORD | The list of all issues of the conversation. |
| issues.name | STRING | The name of the issue. |
| issues.score | FLOAT | Score indicating the likelihood of the issue assignment, between 0 and 1.0. |
| entities | REPEATED RECORD | The list of all spoken entities. |
| entities.name | STRING | Name of the entity. |
| entities.type | STRING | Type of the entity. |
| entities.sentimentScore | FLOAT | The entity sentiment score between -1.0 (negative) and 1.0 (positive). |
| entities.sentimentMagnitude | FLOAT | A non-negative number from zero to infinity that represents the abolute magnitude of the entity sentiment regardless of score. |
| entities.salience | FLOAT | Salience score of the entity. |
| entities.speakerTag | INTEGER | The speaker that the entity mention originated from. |
| labels | REPEATED RECORD | The list of all user-provided labels. |
| labels.key | STRING | User-provided label key. |
| labels.value | STRING | User-provided label value. |
| words | REPEATED RECORD | The list of all spoken words. |
| words.word | STRING | The transcribed word. |
| words.startOffsetNanos | INTEGER | Time offset in nanoseconds of the start of this word relative to the beginning of the conversation. |
| words.endOffsetNanos | INTEGER | Time offset in nanoseconds of the end of this word relative to the beginning of the conversation. |
| words.speakerTag | INTEGER | The speaker that the word originated from. |
| words.languageCode | STRING | Language code. |
| sentences | REPEATED RECORD | The list of conversation sentences. |
| sentences.sentence | STRING | The transcribed sentence. |
| sentences.sentimentScore | FLOAT | The sentence sentiment score between -1.0 (negative) and 1.0 (positive). |
| sentences.sentimentMagnitude | FLOAT | A non-negative number from zero to infinity that represents the abolute magnitude of the sentence sentiment regardless of score. |
| sentences.speakerTag | INTEGER | The speaker that the sentence originated from. |
| sentences.languageCode | STRING | Language code. |
| sentences.startOffsetNanos | INTEGER | Time offset in nanoseconds of the start of this sentence relative to the beginning of the conversation. |
| sentences.endOffsetNanos | INTEGER | Time offset in nanoseconds of the end of this sentence relative to the beginning of the conversation. |
| sentences.createTimeNanos | INTEGER | Time in nanoseconds that the conversation message took place, if provided. |
| sentences.participantId | STRING | Participant ID, if provided. |
| sentences.participantRole | STRING | Participant role, if provided. |
| sentences.obfuscatedExternalUserId | STRING | Obfuscated external user ID, if provided. |
| sentences.isCoveredBySmartReplyAllowlist | BOOL | Whether the sentence is covered by a configured smart reply allowlist in Agent Assist. |
| sentences.annotations | REPEATED RECORD | The list of annotations from Dialogflow Runtime Integration. See full list. |
| sentences.annotations.record | STRING | The annotation record returned from Dialogflow, serialized as JSON. |
| sentences.annotations.createTimeNanos | INTEGER | The timestamp when the suggestion was generated. |
| sentences.annotations.type | STRING | The type of suggestion. |
| sentences.annotations.clicked | BOOL | Customer feedback on whether the suggestion was clicked. |
| sentences.annotations.correctnessLevel | STRING | Customer feedback on the correctness level of the suggestion. |
| sentences.annotations.displayed | BOOL | Customer feedback on whether the suggestion was displayed. |
| sentences.intentMatchData | REPEATED RECORD | The list of matched intents for the sentence. |
| sentences.intentMatchData.intentId | STRING | The unique ID of the matched intent. |
| sentences.intentMatchData.displayName | STRING | The human readable name of the matched intent. |
| sentences.phraseMatchData | REPEATED RECORD | The list of phrase match data for the sentence. |
| sentences.phraseMatchData.phraseMatcherId | STRING | The unique ID of the phrase matcher. |
| sentences.phraseMatchData.revisionId | STRING | Indicating the revision of the phrase matcher. |
| sentences.phraseMatchData.displayName | STRING | The human readable name of the phrase matcher. |
| sentences.highlightData | REPEATED RECORD | The list of highlighter data for the sentence. |
| sentences.highlightData.highlighterName | STRING | The unique id of the highlighter. |
| sentences.highlightData.displayName | STRING | The human readable name of the highlighter. |
| sentences.highlightData.type | STRING | The type of the highlighter. |
| sentences.dialogflowIntentMatchData | REPEATED RECORD | The list of Dialogflow intent match data for the sentence. |
| sentences.dialogflowIntentMatchData.intentName | STRING | The resource name of the matched intent. |
| sentences.dialogflowIntentMatchData.displayName | STRING | The human readable name of the matched intent. |
| sentences.dialogflowIntentMatchData.maxConfidence | FLOAT | The confidence score for the intent match, between 0 and 1.0. |