Skip to content

Feature: GERDA Dispatch Engine

GERDA (**G**rouping, **E**valuation, **R**anking, and **D**ispatch **A**lgorithm) is the intelligent core of Ticket Masala, responsible for multi-factor work assignment and workload optimization.

The Innovation: Algorithmic Dispatching

Most ticketing systems rely on "Round Robin" or manual assignment. Ticket Masala uses a sophisticated scoring engine that evaluates five key dimensions for every assignment:

  1. WSJF Priority: Weighted Shortest Job First ensures high-value, small-effort items are handled first.
  2. Affinity Routing: Connects customers with agents they've worked with previously.
  3. Skill Matching: Verifies agent proficiency level against the specific requirements of the work item.
  4. Workload Balancing: Actively penalizes assignments to agents who are already at high capacity.
  5. Explainability: Generates human-readable reasoning (e.g., "+85 Skill Match, -20 Workload") for every recommendation.

Business Value

The Problem: Manual Bottlenecks

Team leads spend hours every morning manually assigning work from Excel exports. This results in skill mismatches, agent burnout, and inconsistent customer service.

The Solution: Optimized Workforce

GERDA automates the "Matchmaking" between work and agents, improving throughput by 30% and significantly reducing SLA breaches.

Technical Architecture

graph TD
    subgraph "Ingestion"
        New[New Item] -->|Classify| Classifier[ML.NET Classifier]
    end

    subgraph "GERDA Core"
        Classifier -->|Score| Master[GERDA Algorithm]
        Master -->|Affinity| History[(Work History)]
        Master -->|Skills| AgentSkills[(Skills DB)]
        Master -->|Load| CurrentLoad[(Load Monitor)]
    end

    subgraph "Human Review"
        Master -->|Suggest| UI[Atelier Dashboard]
        UI -->|Confirm| Assign[Assign to Agent]
    end

Detailed Capabilities

1. WSJF Scorer

The system calculates a priority score based on: - Business Value: (e.g., Potential Tax Revenue / Debt) - Time Criticality: (SLA urgency) - Risk Reduction: (Detection of Fraud/Sensitive patterns) - Job Size: Predicted effort from ML.NET.

2. The Master Scoring Function

For every candidate agent, GERDA calculates a weighted total: 

// Simplified Scoring Logic
float score = 0;
score += CalculateSkillMatch(agent, item);    // 0-100 pts
score += GetAffinityBonus(agent, customer);  // +50 pts
score -= GetWorkloadPenalty(agent);          // Up to -250 pts
score += agent.YearsOfExperience * 2;        // Seniority bonus

3. Dispatch Explainability

Every recommendation is stored with a payload explaining the factors involved. This transparency builds trust with agents and allows them to understand why they were chosen for a specific case.

Operational Scenarios

Prevent High-Performer Burnout

If Agent Marie is at 90% capacity, GERDA will automatically steer even "perfect match" cases toward Agent Jean (who is at 50% capacity), even if Jean has slightly lower skill proficiency. This ensures long-term team sustainability.

Continuity Wins

A customer calling three times over two weeks will be routed to the same agent 70% of the time, eliminating the frustration of having to repeat the same story to multiple people.

Success Criteria

  1. Acceptance Rate: 80%+ of GERDA suggestions are accepted by team leads.
  2. Balance: No agent exceeds 90% utilization while others are below 70%.
  3. SLA: 95% of "Urgent" cases assigned within 1 hour.

References