Data Source: AI Agents

Data Source: AI Agents

Overview

The Logic+ AI Agent data source provides interaction records generated by AI Agent conversations, including call details, AI-generated summaries, sentiment and topic data, tool call activity, and escalation outcomes. It is useful for reporting on AI Agent performance, self-service rates, and conversation quality.

As with other Insights data sources, the fields below can be used to build reports, create derived fields, apply filters, set groupings, and build charts.


Field Reference

Field Name

Type

Description

Field Name

Type

Description

abandon_reason

String

Reason the call was abandoned (for example, caller hung up or no answer). Primarily relevant for incomplete or abandoned interactions.

agent_id

String

The unique identifier for the AI Agent configuration used for the interaction.

agent_name

String

The human-readable name of the AI Agent as configured.

call_direction

String

Indicates whether the interaction was inbound or outbound. Values: inbound, outbound.

call_duration

Integer

Total call duration in seconds from start to end.

call_rating

String

Call quality or experience rating, if configured. Values may include: positive, neutral, negative.

call_successful

Boolean

Whether the call completed successfully per the AI Agent. This reflects the conversation's own success flag, not a telephony-level record.

call_topic

String

The primary topic of the conversation as identified during the interaction (for example, billing_inquiry, cancellation).

called_number

String

The phone number dialed — the destination number for the call.

caller_frustration_level

String

A qualitative assessment of caller frustration during the interaction. Values may include: low, medium, high.

caller_number

String

The caller's originating phone number.

caller_sentiment

String

Overall sentiment for the caller's side of the conversation. Values may include: positive, neutral, negative.

conversation_id

String

The unique AI Agent conversation identifier. This is the primary key for the interaction at the provider level.

crm_companyId

String

The identifier for the customer's account or company record in a connected CRM, when populated via integration or dynamic variables.

custom_data_points

JSON Object

All custom dynamic variables captured during the interaction. This field often contains customer-specific metrics or identifiers set during the conversation.

custom_llm_response

Text

Additional AI-generated analysis or output configured in the AI Agent's prompts, beyond the core summary.

end_time

Timestamp

Call completion timestamp in ISO 8601 format.

error_code

String

Technical error code for a failed call initiation (for example, NO_ANSWER).

failure_reason

String

Human-readable explanation for a call initiation failure.

has_audio

Boolean

Indicates that a conversation-level audio recording exists for the call.

has_response_audio

Boolean

Indicates that the AI Agent's response audio is available.

has_user_audio

Boolean

Indicates that the caller's audio is available.

interaction_id

String

Internal unique identifier for the AI Agent record within Insights.

next_actions

Text

Recommended follow-up tasks or actions derived from the conversation.

organization_id

String

Internal identifier for the customer organization or account associated with the interaction.

path_taken

String

Identifier of the conversation flow or route taken through the AI Agent (for example, which dialog path or scenario was followed).

recording_url

String

URL where the final audio recording for the interaction is accessible.

retry_attempts

Integer

Number of attempts to initiate the call before a failure was finalized.

self_serve_status

Boolean

Indicates whether the customer was fully self-served by the AI Agent without being transferred to a human agent.

speaking_breakdown

String

Text representation of speaking-time distribution between the caller and the AI Agent (for example, caller: 60%, agent: 40%).

start_time

Timestamp

Call initiation timestamp in ISO 8601 format.

summary

Text

AI-generated summary describing the full interaction.

tool_calls

JSON Array

Detailed list of tool invocations during the interaction, including tool name, parameters, results, timestamps, and success flags.

tool_results

JSON

Aggregated or extracted results from tool executions during the interaction.

transfer_required

Boolean

Indicates whether the interaction required escalation or transfer to a human agent.

user_id

String

A customer-provided user identifier associated with the interaction, if supplied to the AI Agent.

Related Articles