> ## Documentation Index > Fetch the complete documentation index at: https://koreai.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Creating Audit Allocations Creating Audit Allocations provides a guided workflow to create manual audit allocations for Human Agent and AI Agent conversations. QA Managers can define the evaluation scope, select interactions using Random or Custom allocation methods, distribute audits across evaluators, and configure dispute handling for Human Agent evaluations. The workflow automatically adapts to the selected evaluation form, displaying only the applicable configuration options. ## Access Audit Allocations Navigate to **Quality AI** > **AutoQA** > **Allocations** > **New Audit Allocation**. Create Allocations *** ### Step 1: Settings Configure the allocation details and define the interactions to evaluate. #### Configure Basic Information | Field | Description | | ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | **Name** | Descriptive name that identifies the allocation purpose and makes it easy to locate later. | | **Description** | (Optional) Describe the allocation purpose or evaluation focus, or evaluation priorities to help team members understand the intent. | | **Date Range** | Specify the date range for interactions to evaluate and capture the relevant conversation window. | | **Evaluation Form** | Select an evaluation form to guide auditors in reviewing and assessing assigned interactions. This determines the linked queues available for filtering and defines the dispute workflow, including routing rules, re-dispute limits, and dispute settings for evaluations under the allocation. | | **Languages** | Select one or more languages. Select specific languages when creating audits for auditors who evaluate only certain languages, or select all languages for comprehensive coverage. The interaction count updates automatically to reflect the selection. | The selected evaluation form determines the available queues and interaction directions, including Inbound, Outbound, or Both. This supports both customer-initiated and agent-initiated interactions, such as callbacks and campaigns. #### Filter by Queue, Agents or Agent Groups Limit the allocation to specific queues, agents or agent groups. | Filter | How to use | | ----------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Queue** | Queue associated with the selected evaluation form. | | **Agents** | Select individual agents from the dropdown. Leave empty to include all agents in the selected queues. Users with the Cross Queue Data Access permission can select agents across all queues. | | **Agent Groups** | Select pre-configured groups for bulk selection. Useful for team-based evaluations. Combine with individual agent selection as needed. | | **Selected Interactions for Audit** | Displays the total number of interactions that match the selected filters and are eligible for allocation, along with a breakdown of eligible Voice and Chat interactions. | The displayed counts are estimates based on the current selection criteria and may change if the allocation settings or interaction data are updated before the allocation is created. Select **Next** to proceed to the **Allocation** step. Settings *** ### Step 2: Allocation Type Choose how interactions are selected from the filtered dataset. #### Random Allocation Uses statistical sampling to ensure impartial selection of interactions, providing representative coverage across agents with control over sample size. 1. Select the **Random** option. 2. Enable the **Channel Type**: **Voice**, **Chat**, and **Both**. 3. (Optional) Enable **Include interactions below duration threshold** to include interactions that fall below the configured minimum duration threshold. * **Enabled** (default)\*\*: Below-threshold interactions are excluded from random sampling. * **Disabled**: The system includes below-threshold interactions in the allocation and displays a Below Threshold indicator during evaluation. 4. Configure separate allocation logic for **Voice** and **Chat** interactions. | Logic | Description | | -------------------------- | -------------------------------------------------------- | | **Percentage Per Agent** | Allocates a percentage of interactions for each agent. | | **Interactions Per Agent** | Allocates a fixed number of interactions for each agent. | 5. Enter the required percentage or interaction count for each enabled channel. The system automatically displays allocation metrics such as average selected interactions for Voice, average interactions per agent for Chat, and the total selected interactions for both Voice and Chat. 6. Review the audit summary, which includes the total number of selected interactions and the breakdown of Voice and Chat interactions. This setting is configurable independently for Voice and Chat channels, enabling flexible allocation behavior across different interaction types. Allocation Type #### Custom Allocation Applies specific filters to select interactions based on scenarios, sentiment ranges, or performance metrics. Ideal for focused evaluations or validating AutoQA accuracy. 1. Select the **Custom** option. 2. Choose one of the following filter sources. * **Saved Filter**: Select a filter previously saved in Conversation Mining. Best for recurring audit focuses or standard evaluation scenarios. * **Quick Filter**: Configure a filter directly in the allocation workflow. Best for one-time allocations without saving filters. Perfect for one-time allocations without cluttering your saved filter library. Custom Allocation Select at least one interaction for audit allocation to proceed. **Quick Filter Creation** Select **Create Filter** to open the quick filter builder. | Filter Category | What you can filter | | --------------- | ------------------------------------------------------------------------------------------------ | | **Experience** | Sentiment trends, emotions, wait time, CSAT, intents, topics, churn risk. | | **Behavior** | Evaluation metrics, empathy score, crutch word usage, playbook adherence, Kore Evaluation Score. | | **Keyword** | Include or exclude interactions containing specific words or phrases. | | **Efficiency** | AHT, handling time ranges, AHT deviation, number of transfers. | The interaction count updates dynamically as you adjust filters. For more information on filter interactions, see [Filter Interactions](/ai-for-service/quality-ai/analyze/conversation-mining-interactions#filter-interactions). Quick Filter Creation **Custom Interaction Allocation** After you define the filters, configure interaction allocation settings to determine how the system assigns interactions for audit. * **All Interactions**: Includes all interactions returned by the applied filters. * **Custom**: Enables you to apply additional sampling controls to the filtered interaction set. When you select **Custom**, choose one of the following allocation methods: | Allocation Method | Description | | -------------------------- | ----------------------------------------------------------------- | | **Percentage Per Agent** | Allocates a percentage of filtered interactions for each agent. | | **Interactions Per Agent** | Allocates a fixed number of filtered interactions for each agent. | The system automatically displays the average number of selected interactions based on your configuration. **Example**: Filter interactions with sentiment scores lower than 3, then allocate 5% per agent from those filtered interactions to focus on low-satisfaction conversations. Select **Next** to proceed to the **Assignment** step. *** ### Step 3: Assignment Distribute selected interactions across the auditor team, ensuring balanced workload and activating the allocation for evaluation. 1. Use the **Search Auditors** dropdown to select team members who evaluate interactions. 2. Choose multiple auditors to distribute the workload. Consider these factors when selecting auditors. 3. Enter the **Allocation Percentage** for each auditor. The system calculates the exact interaction count per auditor automatically. 4. Review the **Interactions** column, which automatically calculates the exact interaction count for each auditor based on your percentage input. **Example** (100 total interactions, 3 auditors): * Auditor A: 40% = 40 interactions * Auditor B: 30% = 30 interactions * Auditor C: 30% = 30 interactions 5. Verify that the total equals 100%. 6. Review your complete allocation configuration: * **Settings**: Form, languages, agents, agent groups. * **Allocation**: Method (Random or Custom), filters, total interactions. * **Assignment**: Auditors and their percentages. 7. Select **Save** to activate the allocation. When activated, the system immediately assigns interactions to auditors. Auditors can see their assignments in the **Assigned to Me** tab right away. Assignment #### After Creating an Allocation * The allocation displays in your **My Allocations** tab. * Auditors see their assigned interactions in **Assigned to Me**. * Track completion progress in real-time. * AI Agent evaluations are processed as system-scoped evaluation sets. * Edit the allocation at any time to reassign pending interactions. * You can reassign the pending interactions at any time without affecting completed evaluations. The system automatically calculates the interaction count for each auditor and enables **Save** only when the allocation totals 100%. If you enable Agent Accept & Dispute and configure disputes for the selected evaluation form, the Review Status column tracks agent acknowledgments and dispute activity for all audited interactions in the allocation.> *** ### Step 4: Dispute Allocation (Human Agent Only) Available only when **Agent Accept & Dispute** is enabled. #### Dispute Resolution Assignment The **Dispute Resolution Assignment** setting to enable or disable disputes feature for evaluations under this allocation. | State | Behavior | | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------ | | **On** | Enable agents to acknowledge or dispute metric-level scores for evaluations under this allocation. | | **Off** | Prevent agents from responding to evaluations. The system hides the **Accept** and **Dispute** options and marks evaluations as **Audited** upon completion. | #### Dispute Routing Rule When you enable **Dispute Resolution Assignment**, select how the system routes disputes for re-evaluation. | Option | Description | | --------------------- | --------------------------------------------------------------------- | | **Same Auditor** | Route disputes to the auditor who performed the original evaluation. | | **Different Auditor** | Route disputes to a different auditor selected from the auditor list. | #### Auditor List (Different Auditor Only) When you select **Different Auditor**, define the pool of auditors available for dispute routing. **Add Auditors** * Select **Search Auditors to Add**. * Search for and select one or more auditors. * Select **Add**. Added auditors display in the list below the search field. If you don't add any auditors, the system shows **No Auditors Added**. #### Allow Multi-Round Re-Disputes The **Allow Multi-Round Re-Disputes** option controls whether agents can raise additional disputes after QA re-evaluation. | State | Behavior | | ------- | ----------------------------------------------------------------------------------------------------------- | | **On** | Enable agents to dispute QA re-evaluations and initiate additional rounds up to the configured limit. | | **Off** | Make the QA decision final. The system prevents further disputes and shows an in-context message to agents. | #### Maximum Re-Disputes Allowed When you enable multi-round disputes, select the maximum number of dispute rounds. | Setting | Description | | ------------------------------- | ------------------------------------------------------------------- | | **Maximum Re-Disputes Allowed** | Select the number of dispute rounds from the dropdown (default: 3). | When the system reaches the limit, it changes the evaluation status to **Closed** and prevents any further disputes. Dispute configuration is not applicable for AI Agent allocations. You can edit an allocation at any time to redistribute pending interactions. Completed evaluations remain unchanged. ## Use Case Example: Redistribute Pending Audits **Scenario**: A QA Manager creates a monthly random sample allocation for 15 auditors. Three days later, two auditors report unexpected leave, leaving 40 pending interactions unassigned. ### Scenario A QA Manager creates a monthly audit allocation for multiple auditors. During the audit cycle, two auditors become unavailable, leaving several pending interactions unreviewed. ### Resolution The QA Manager edits the allocation, selects the unavailable auditors, and reassigns their pending interactions to available auditors by updating the workload distribution. The system redistributes only the pending interactions, while completed evaluations remain unchanged. Newly assigned auditors immediately see the reassigned interactions in Assigned to Me. ### Outcome The system keeps the monthly audit on schedule, prevents loss or duplication of interactions, and balances workload across available team members. ***