Conversation Sources lets you configure and manage voice and chat conversation inputs from multiple platforms in one place. Quality AI supports ingestion from Contact Center AI (CCAI), Agent AI, and Quality AI Express.
You can assign agents to queues directly through the UI, enable or disable sources, and view and manage queues — including queue name, queue ID, system-generated ID, and assigned agents.
Access Conversation Sources
Navigate to Quality AI > CONFIGURE > Conversation Sources.
Supported Sources
Quality AI supports three conversation sources:
Contact Center AI (CCAI)
Ingests real-time conversation data from the CCAI platform.
| Property | Details |
|---|
| Integration method | Direct API connection |
| Data processing | Real-time |
| Configuration | None required (automatic) |
| Use case | Organizations using Kore.ai’s native contact center |
If CCAI is disabled, the system stops processing interactions from CCAI on third-party desktops. CCAI queues do not support agent-queue mapping through the UI — queue configuration follows existing CCAI workflows and APIs.
Agent AI
Supports human-agent interactions enhanced by AI.
| Property | Details |
|---|
| Integration method | Structured data feeds |
| Data processing | Real-time with metadata validation |
| Configuration | Add queues with Agent Queue names and Queue IDs |
| Use case | Organizations with AI-assisted human agents |
If Agent AI is disabled, the system stops processing interactions from Agent AI on third-party desktops. Include agent email IDs and queue IDs during integration.
Quality AI Express
Imports interactions from external sources (service records, chat logs, emails) using AWS S3 Connectors.
| Property | Details |
|---|
| Integration method | AWS S3 Connectors |
| Data processing | Batch file processing |
| Configuration | Add queues with agent queue names and queue IDs |
| Supported data | Service records, chat logs, emails, audio recordings, metadata files |
| Use case | Organizations using third-party CCaaS platforms |
When you enable Quality AI Express, the system displays Connectors in the Quality AI left navigation panel. Connectors are the primary bridge for importing conversations through AWS S3 file transfers.
Before You Configure
Complete these steps before activating sources:
- Assign appropriate user permissions.
- Verify required metadata structures exist.
- Confirm integration endpoints are available.
Quality AI Express Prerequisites
Before activating Quality AI Express:
- Clear all active Playbook metrics from evaluated form queues.
- Clear all active Dialog Task metrics from evaluated form queues.
If you skip these steps, the system blocks activation and displays a warning.
Also configure:
- S3 bucket access
- File transfer protocols
- Data format specifications
- Processing schedules
Interface Overview
The Conversation Sources interface shows all enabled sources and their queues.
| Element | Description |
|---|
| Enable/Disable toggles | Start or stop ingestion per source |
| Queue list | Shows queue name, Queue ID, system-generated ID, and agent count |
| Actions | Edit or delete queues |
| Add Queue | Create new queues and assign agents |
Agent assignments made through public APIs may take time to appear in the UI. Refresh or re-sync queues to see updated mappings.
Queue Management
Configure queues for Agent AI and Quality AI Express to route conversations correctly.
Add a Queue
-
Navigate to Quality AI > CONFIGURE > Conversation Sources.
-
Enable Agent AI, Quality AI Express, or both.
-
Select Add Queue and fill in the details:
| Field | Description |
|---|
| Name | Descriptive name for the queue |
| Queue ID | Unique identifier for the queue |
| System Generated Queue ID | Auto-generated ID for internal tracking |
| Agents | Assign agents using the searchable list |
-
Select Save.
Agent–Queue Mapping
Direct agent-to-queue mapping is available for Agent AI and Quality AI Express queues through the UI.
For interactions to appear in the system, all three conditions must be met:
- The queue is mapped to an evaluation form.
- The agent is mapped to the queue through the UI.
- The supervisor or agent viewing the data belongs to the same queue.
If any condition is not met, interactions will not appear.
CCAI queue configuration and agent assignment are not affected by UI-based mapping. You can assign agents created through platform-level user administration and public APIs.
Chat Script Timestamp Format
This setting controls how Quality AI Express parses timestamps in chat conversation files during ingestion. The selected format applies to all ingested chat files and persists across sessions.
Why timestamp format matters:
- Orders messages chronologically
- Calculates conversation duration and response times
- Identifies hold periods and transfer events
- Generates time-based analytics
Unix Timestamp
Represents seconds since the Unix epoch.
- Format: Integer or decimal number
- Examples:
1735574400 or 1735574400.523
The ingestion process validates start and end times, which may limit flexibility for demo environments and third-party desktop integrations.
Offset Timestamp
Represents seconds from the start of the conversation.
- Format: Integer or decimal number
- Examples:
0 (start), 45 (45 seconds in), 120.5 (2 min 0.5 sec in)
This format does not validate start and end dates — timestamps are used only for reporting and filtering. This change applies only to chat ingestion.
Source Disable Warnings
| Source | Effect of disabling |
|---|
| CCAI | Stops processing CCAI interactions |
| Agent AI | Stops processing Agent AI interactions |
| Quality AI Express | Blocks file-based ingestion |
Access Control
| Role | Permissions |
|---|
| Auto QA access | Enable/disable Quality AI Express, save and modify settings |
| Limited access users | View-only interface, cannot save configurations |
| Agent-level access | Toggle enable/disable, cannot save settings |
When you disable Quality AI Express, the following CCAI metrics are hidden:
- Compare functionality
- Contact Center Efficiency
- Agent Efficacy (including CSAT)
- Agent Performance Monitor (Occupancy, Playbook Adherence)
- Customer Experience (Average Wait Time, NPS Score)
- Interaction Details (Dispositions)