AI Audit

The AI Audit module provides visibility into all AI-driven operations performed within the system. It captures detailed logs of AI feature usage, including the model used, execution time, request and response data, token consumption, and status.

What you see on the screen

The AI Audit page displays a consolidated list of all AI interactions performed across the platform. Each record represents a single AI request, including its execution details, such as feature, model, time taken, and response status.

Users can search, filter, and analyze audit records using multiple parameters to identify specific AI operations or performance insights.

AI Audit | Basic Details

Label
Action
Description / Example

Search

Input text

Allows users to search audit records using keywords such as feature name, message content, or user.

Filter

Apply conditions

Enables filtering of audit data based on selected fields and conditions. Available Fields: Feature, Model, Request Time, Message, Response Time, Status, Settings, User, Provider, Request Token, and Response Token.

Date Range

Select from the dropdown

Filters audit records based on a selected time period (e.g., Last 60 Days).

Audit Data Fields | AI Audit

The following columns are available in the audit table:

Label
Action
Description / Example

Feature

View

Displays the AI feature used (e.g., KB Creation, Service Classification).

Model

View

Shows the AI model used for execution (e.g., GPT-5.4-mini).

Request Time

View

Indicates when the AI request was initiated.

Response Time

View

Indicates when the AI response was generated.

Time Taken (sec)

View

Total execution time for the request.

Request Token

View

Number of tokens used in the request.

Response Token

View

Number of tokens generated in the response.

Total Token

View

Combined token usage (request + response).

Message

View

Displays the input sent to the AI model.

Response Message

View

Displays the AI model's output.

Status

View

Shows execution status (e.g., success, error).

LLM Settings

View

Displays configuration details used for the AI request.

User

View

Indicates the user/system that triggered the AI action.

Error

View

Displays error details if the execution fails.

Key Type

View

Indicates the type of API key used (e.g., External).

Key Name

View

Displays the configured key name.

Provider

View

Shows the AI provider (e.g., OpenAI, Azure).

Last updated

Was this helpful?