# Core Concepts ### Review Kinds Review Kinds define the different types of reviews your team performs. Think of them as templates that specify what type of review is needed and how urgently it should be completed. Each Review Kind includes: * **Name & Description**: Clear identification and instructions that help reviewers understand what they're evaluating * **Base Priority** (1-10 scale): Starting priority level that determines the review's initial importance relative to other review types * **SLA Timeline**: Hours from creation until the review's due date * **Escalation Parameters**: Controls how the review's priority increases as it approaches its deadline By configuring different Review Kinds, you can ensure that time-sensitive fraud investigations receive appropriate urgency compared to routine document checks. *** ### Review Queues Review Queues organize reviews for different teams or workflows. While Review Kinds define *what* needs to be reviewed, Queues determine *who* reviews it and *how* work is prioritized within that team. **Key capabilities of Review Queues:** * **Team Segmentation**: Create separate queues for different departments (e.g., Compliance Team, Fraud Specialists, L1 Reviewers) * **Priority Configuration**: Choose how reviews are ordered within each queue using different sorting strategies * **Review Kind Binding**: Specify which types of reviews should flow into each queue * **Review Agent Assignment**: Control which team members have access to pull reviews from each queue **Example queue structure:** * **General Queue**: Handles standard document verifications and low-risk reviews for your L1 team * **Escalations Queue**: Receives cases that AI agents couldn't resolve or that reviewers manually escalated * **Specialist Compliance Queue**: Routes high-risk, complex cases to your most experienced reviewers This segmentation ensures that your specialized resources focus on cases that truly need their expertise, while routine reviews move efficiently through your general team. *** ### Queue Prioritization Strategies Understanding how reviews are ordered within each queue is critical for operational efficiency. The Review Platform offers four distinct sorting strategies, each designed for different operational needs. ### SLA Sort Orders reviews by earliest due date first. This strategy ensures you never miss a deadline but doesn't account for the relative importance of different review types. **Best for**: Teams where all review types have similar importance and deadline compliance is the primary concern. ### Relative Priority Sort Orders reviews by highest base priority first, regardless of when they were created or when they're due. **Best for**: Teams handling both routine and critical reviews where certain types (like fraud investigations) should always take precedence. ### Created Sort Orders reviews by most recent first, functioning as a simple FIFO (first-in, first-out) queue. **Best for**: Teams with consistent review types where age alone is the best indicator of urgency. ### Hybrid Prioritization (Recommended) Dynamically calculates priority using multiple factors simultaneously: base priority level, proximity to SLA deadline, and creation time. This is the most sophisticated option and reflects real operational complexity. **How Hybrid Prioritization works:** The system continuously recalculates each review's effective priority using this formula: ``` Priority = Base Priority × (1 + (Max Multiplier - 1) × Progress^SLA Ramp Factor) ``` **Parameters:** * **Base Priority**: The review kind's starting priority (1-10) * **Max Multiplier**: How much priority can increase (e.g., 2x means priority can double) * **Progress**: Percentage of time elapsed toward SLA deadline (0-100%) * **SLA Ramp Factor**: Controls the escalation speed **SLA Ramp Factor options:** * **Linear (1.0)**: Steady, predictable escalation over time. A review's priority increases proportionally as it approaches its deadline. * **Exponential (>1.0)**: Slow escalation early, then rapid increase near deadline. Use this when you want to give newer reviews time to be addressed before older reviews become urgent. * **Logarithmic (\<1.0)**: Fast escalation early, then gradual increase. Use this when you want older reviews to quickly gain priority. **Why Hybrid Prioritization matters:** Consider two reviews in your queue with identical basic metrics: **Review A: High-Value Transaction** * Base priority: 5 * SLA: 4 hours * Created: 3.5 hours ago * $50,000 wire transfer flagged for manual review * Customer is a VIP with urgent business needs **Review B: Standard Transaction** * Base priority: 5 * SLA: 4 hours * Created: 3.5 hours ago * $500 credit card transaction flagged for manual review * Routine verification needed ### Hybrid Prioritization Advantage With **simple sorting strategies**, both reviews would be treated identically, where you'd ideally want to review the high value transaction first: * **SLA sorting**: Both reviews have the same time due * **Review creation sorting**: Both reviews have the same creation time * **Relative priority sorting**: Both reviews have the same relative priority With **Hybrid Prioritization**, the system can incorporate additional contextual factors to make sure important reviews get done first *** ## Reviewers and Teams Overview The Review Platform supports both human reviewers and AI agents working together to handle your onboarding workflows efficiently. ### Human Review Agents **Review Agents** are your team members who perform manual reviews. The platform provides comprehensive management capabilities: * **Queue Access Control**: Assign agents to specific queues based on their expertise and specialization * **Performance Tracking**: Monitor individual productivity, review quality, and throughput metrics * **Daily Quotas**: Set realistic targets to help manage workload and expectations * **Permission Levels**: Control what actions agents can perform (review, escalate, approve, etc.) ### Review Teams **Review Teams** organize agents for better management and operational efficiency: * **Specialization Groups**: Create teams like Fraud Specialists, Document Reviewers, or Compliance Teams * **Performance Comparison**: Track and compare metrics across different operational groups * **Escalation Management**: Define clear paths for complex cases that need expert attention * **Workload Distribution**: Ensure equitable work distribution across team members ### AI Agents Integration The platform seamlessly integrates with AI agents that can handle routine verifications and automated decisioning: * **Automated Processing**: AI agents can resolve straightforward cases without human intervention * **Escalation Routing**: Cases that AI agents cannot resolve automatically flow to appropriate human review queues * **Hybrid Workflows**: Combine AI efficiency with human expertise for optimal coverage * **Quality Assurance**: Human reviewers can validate AI decisions and provide feedback for continuous improvement This multi-agent approach ensures that simple cases are processed quickly by AI while complex or uncertain cases receive the attention of your specialized human reviewers.