ANSWERING schema reference

This is for:

In this article

ANSWERS
ANSWER_INTERACTIONS
CONFIGS
CONVERSATIONS

This article provides a reference to the views found in the ANSWERING schema of the CORE data model. The ANSWERING schema captures data for both Relevance Generative Answering (RGA) and Search Agent, with a focus on generated answers, conversations, agent configurations, and user interactions.

ANSWERS

The ANSWERS view consists of comprehensive details about generated answers within an organization, covering both Relevance Generative Answering (RGA) and Search Agent turns. Each row represents a unique answer or turn, indicates whether a response was generated, and includes information about its origin, generation time, and key identifiers.

Column name Description

Column name	Description
`ANSWER`	The content of the generated answer.
`ANSWER_DATE`	UTC date when the answer occurred.
`ANSWER_ID`	Unique identifier of this answer, used as primary key. References `INSIGHT_ID` in the `INSIGHTS` view. Each answer is modeled as an insight and this view provides extended information specific to answers.
`ANSWER_LENGTH`	Length of the answer in characters.
`ANSWER_SOURCE`	The source of the answer. Possible values are `search_agent`, `crga`, or `NULL`. A `NULL` value indicates that the row was created before this column was added to the view.
`ANSWER_TIMESTAMP`	UTC timestamp when the answer occurred.
`ANSWER_XID`	External identifier for this answer. For RGA answers, this is the platform’s `answerid`. For Search Agent turns, this is the Agent API’s `turnId`.
`CONFIG_ID`	Unique identifier of the configuration that produced this answer. Keys into `CONFIGS`.
`CONVERSATION_ID`	Unique identifier of the conversation this answer belongs to. Keys into `CONVERSATIONS`. The value is `NULL` for RGA answers.
`IS_COMPLETELY_STREAMED`	Boolean indicating whether the answer was completely streamed.
`IS_FOLLOWUP_ENABLED`	Boolean indicating whether follow-up turns are enabled for this answer. The value is `NULL` for RGA answers.
`IS_QUERY_ANSWERED`	Boolean indicating whether the answer was generated.
`IS_RECEIVED`	Boolean indicating whether the answer was received by the client.
`ORGANIZATION_ID`	Organization identifier for the record.
`PARENT_ANSWER_ID`	The `ANSWER_ID` of the initial turn in the conversation. The value is `NULL` for first turns and for RGA answers. Keys into `ANSWERS`.
`RELATED_INSIGHT_ID`	Identifier of a related insight. Used when the answer is linked to another insight (for example, when a search and an answer are returned for the same query). Keys into `INSIGHTS`.
`TURN_INDEX`	1-based position of this turn within the conversation.
`VISIT_ID`	Reference to the visit in which this answer happened. Keys into `VISITS`.

ANSWER

The content of the generated answer.

ANSWER_DATE

UTC date when the answer occurred.

ANSWER_ID

Unique identifier of this answer, used as primary key. References INSIGHT_ID in the INSIGHTS view. Each answer is modeled as an insight and this view provides extended information specific to answers.

ANSWER_LENGTH

Length of the answer in characters.

ANSWER_SOURCE

The source of the answer. Possible values are search_agent, crga, or NULL. A NULL value indicates that the row was created before this column was added to the view.

ANSWER_TIMESTAMP

UTC timestamp when the answer occurred.

ANSWER_XID

External identifier for this answer. For RGA answers, this is the platform’s answerid. For Search Agent turns, this is the Agent API’s turnId.

CONFIG_ID

Unique identifier of the configuration that produced this answer. Keys into CONFIGS.

CONVERSATION_ID

Unique identifier of the conversation this answer belongs to. Keys into CONVERSATIONS. The value is NULL for RGA answers.

IS_COMPLETELY_STREAMED

Boolean indicating whether the answer was completely streamed.

IS_FOLLOWUP_ENABLED

Boolean indicating whether follow-up turns are enabled for this answer. The value is NULL for RGA answers.

IS_QUERY_ANSWERED

Boolean indicating whether the answer was generated.

IS_RECEIVED

Boolean indicating whether the answer was received by the client.

ORGANIZATION_ID

Organization identifier for the record.

PARENT_ANSWER_ID

The ANSWER_ID of the initial turn in the conversation. The value is NULL for first turns and for RGA answers. Keys into ANSWERS.

RELATED_INSIGHT_ID

Identifier of a related insight. Used when the answer is linked to another insight (for example, when a search and an answer are returned for the same query). Keys into INSIGHTS.

TURN_INDEX

1-based position of this turn within the conversation.

VISIT_ID

Reference to the visit in which this answer happened. Keys into VISITS.

ANSWER_INTERACTIONS

The ANSWER_INTERACTIONS view consists of detailed records of how users interact with specific answers within an organization. Each row represents a unique interaction event that provides insights into user engagement and feedback.

Column name Description

Column name	Description
`ANSWER_ID`	Reference to the related answer record. Keys into `ANSWERS`.
`ANSWER_INTERACTION_ID`	Unique identifier of the answer interaction, used as primary key.
`INTERACTION_TIMESTAMP`	Timestamp of the answer interaction in UTC.
`INTERACTION_DATA`	Contains data specific to the interaction type.
`INTERACTION_TYPE`	The action type performed on the answer, such as `like`, `dislike`, or `expand`.
`ORGANIZATION_ID`	Organization identifier for the record.

ANSWER_ID

Reference to the related answer record. Keys into ANSWERS.

ANSWER_INTERACTION_ID

Unique identifier of the answer interaction, used as primary key.

INTERACTION_TIMESTAMP

Timestamp of the answer interaction in UTC.

INTERACTION_DATA

Contains data specific to the interaction type.

INTERACTION_TYPE

The action type performed on the answer, such as like, dislike, or expand.

ORGANIZATION_ID

Organization identifier for the record.

CONFIGS

The CONFIGS view contains one row per organization and configuration. Each row represents a unique configuration associated with generated answers.

Column name Description

Column name	Description
`CONFIG_ID`	Internally generated unique identifier for the configuration. Used as the primary key.
`CONFIG_XID`	External configuration identifier. For Search Agent answers, this corresponds to the `agentId` from the Agent API. For RGA answers, this corresponds to the answer configuration ID from the Answer API.
`NAME`	Name of the configuration. The value is `NULL` for Search Agent configurations.
`ORGANIZATION_ID`	Organization identifier for the record.

CONFIG_ID

Internally generated unique identifier for the configuration. Used as the primary key.

CONFIG_XID

External configuration identifier.

For Search Agent answers, this corresponds to the agentId from the Agent API.
For RGA answers, this corresponds to the answer configuration ID from the Answer API.

NAME

Name of the configuration. The value is NULL for Search Agent configurations.

ORGANIZATION_ID

Organization identifier for the record.

CONVERSATIONS

The CONVERSATIONS view contains one row per organization and captured conversation. Each row represents a unique Search Agent conversation and its time boundaries, derived from captured turns.

Column name Description

Column name	Description
`CONVERSATION_ID`	Internally generated unique identifier for the conversation. Used as the primary key.
`CONVERSATION_XID`	External conversation identifier, corresponding to the `conversationId` from the Agent API.
`END_DATE`	UTC date of the last turn in the conversation.
`END_TIMESTAMP`	UTC timestamp of the last turn in the conversation.
`ORGANIZATION_ID`	Organization identifier for the record.
`START_DATE`	UTC date of the first turn in the conversation.
`START_TIMESTAMP`	UTC timestamp of the first turn in the conversation.

CONVERSATION_ID

Internally generated unique identifier for the conversation. Used as the primary key.

CONVERSATION_XID

External conversation identifier, corresponding to the conversationId from the Agent API.

END_DATE

UTC date of the last turn in the conversation.

END_TIMESTAMP

UTC timestamp of the last turn in the conversation.

ORGANIZATION_ID

Organization identifier for the record.

START_DATE

UTC date of the first turn in the conversation.

START_TIMESTAMP

UTC timestamp of the first turn in the conversation.