# Predictive Routing

## Study Notes
| Topic | Description |
|---|---|
| Predictive Routing | AI-powered routing system that optimizes agent assignment |
| Engine | Machine learning algorithms analyze skills, availability, and contact history |
| Purpose | Maximize first-contact resolution and customer satisfaction |
| Activation | Requires Premium edition and Workforce Optimization module |
| Benefit | Reduces handle time and improves customer outcomes |

---

## Navigation
Admin → Architect → Routing → Predictive Routing
OR
Admin → Contact Center → Routing Configuration → Enable Predictive Routing

---

## Predictive Routing Overview

Predictive Routing is an AI-powered contact routing system that dynamically matches incoming contacts to the most suitable available agent based on multiple factors:

### Key Capabilities
- **Skill-based routing** - Matches agent skills to contact requirements
- **Historical performance** - Learns from agent interaction outcomes
- **Availability prediction** - Anticipates agent availability and readiness
- **Real-time optimization** - Adjusts routing in real-time based on system state
- **Omnichannel support** - Works across voice, chat, email, and messaging

### How It Works
1. Contact arrives at system
2. Contact intent and requirements analyzed
3. System evaluates all available agents
4. Machine learning algorithm predicts best match
5. Contact routed to optimal agent
6. Interaction data captured for learning

---

## Edition & Module Requirements

| Requirement | Details |
|---|---|
| Minimum Edition | Premium Edition required |
| Module | Workforce Optimization add-on module |
| License Type | Agent licenses with predictive routing enabled |
| Setup | Admin configuration in Architect |

---

## Study Notes - Routing Factors
| Factor | Description | Impact |
|---|---|---|
| Agent Skills | Capabilities and certifications | High - Core matching criteria |
| Proficiency Level | Skill mastery degree | High - Affects quality |
| Availability | Agent ready state and status | High - Real-time factor |
| Handling Capacity | Available slots for new contacts | High - Prevents overload |
| Historical Performance | Past interaction outcomes | Medium - Learning factor |
| Queue Wait Time | Customer wait duration | Medium - Fairness factor |
| Contact Type | Voice, chat, email, etc. | High - Channel match |
| Language Proficiency | Supported languages | High - Communication match |
| Customer History | Previous interaction records | Medium - Context factor |
| Agent State | Idle, working, after-call work | High - Real-time factor |

---

## Implementation Guide

### Step 1: Prerequisites & Planning
1. Ensure organization has Premium edition
2. Purchase Workforce Optimization module
3. Audit existing agent skills database
4. Document required skill sets
5. Review current routing rules
6. Plan migration from legacy routing

### Step 2: Configure Skill Definitions
1. Navigate to Admin → Architect → Skills
2. Create skill categories (technical, language, product)
3. Define skill levels (1-5 proficiency)
4. Assign skills to agents
5. Establish mastery thresholds
6. Document skill requirements per queue

### Step 3: Enable Predictive Routing
1. Go to Admin → Contact Center → Routing
2. Select queue to enable predictive routing
3. Enable "Predictive Routing" toggle
4. Configure routing rules (optional overrides)
5. Set skill matching parameters
6. Define fallback routing behavior

### Step 4: Testing & Validation
1. Route test calls through system
2. Monitor agent assignment accuracy
3. Verify skill matching
4. Check queue distribution
5. Validate omnichannel routing
6. Review abandonment rates

### Step 5: Monitoring & Optimization
1. Review routing analytics daily
2. Monitor first-contact resolution rates
3. Track agent utilization
4. Measure customer satisfaction
5. Optimize skill assignments
6. Adjust thresholds as needed

---

## How to Implement

| Phase | Description | Timeline |
|---|---|---|
| Analysis | Audit current routing and skills | Week 1-2 |
| Configuration | Set up skills, queues, and rules | Week 2-3 |
| Testing | Validate routing logic and assignments | Week 3-4 |
| Pilot | Deploy to test queue with monitoring | Week 4-6 |
| Full Rollout | Enable across all queues | Week 6-8 |
| Optimization | Monitor, tune, and improve | Ongoing |

---

## Predictive Routing Architecture

```
Incoming Contact
    ↓
Contact Metadata Analysis
├── Contact Intent
├── Contact Type (Voice/Chat/Email)
├── Language Requirements
└── Skill Requirements
    ↓
Predictive Routing Engine
├── Machine Learning Models
├── Real-time Agent Analysis
├── Skill Matching Algorithm
└── Performance Prediction
    ↓
Agent Evaluation
├── Available Agents
├── Skill Match Score
├── Proficiency Level
├── Historical Performance
└── Queue Load Balance
    ↓
Optimal Agent Selection
    ↓
Route Contact
    ↓
Agent Assignment
```

---

## Routing Decision Flow

```
Contact Arrives
    ↓
Extract Contact Data
├── Intent
├── Channel Type
├── Language
└── Queue Assignment
    ↓
Predictive Routing Engine Evaluates
├── Agent Availability Status
├── Skill Compatibility
├── Proficiency Levels
├── Current Workload
└── Historical Performance Score
    ↓
Machine Learning Model Predicts
├── Best Agent Match
├── Estimated Resolution Probability
└── Customer Satisfaction Likelihood
    ↓
Route to Optimal Agent
    ↓
If No Optimal Agent Available
├── Queue with Priority Calculation
├── Monitor for Next Available Match
└── Apply Fallback Routing Rules
    ↓
Agent Accepts Contact
```

---

## Routing Rules & Overrides

### Hard Rules (Always Applied)
```
Rule Priority: High
├── Agent Availability
├── Required Skills Present
├── Language Match
└── Queue Assignment

Rule Priority: Medium
├── Skill Proficiency Threshold
├── Agent Capacity
├── Contact Type Capability
└── Channel Configuration

Rule Priority: Low
├── Load Balancing
├── Historical Performance
└── Fairness Rotation
```

---

## Skill Configuration Example

### Technical Support Queue
```
Required Skills:
├── Product Knowledge (Level 3+)
│   ├── Software (Level 4)
│   ├── Hardware (Level 3)
│   └── Cloud Services (Level 3)
├── Troubleshooting (Level 3+)
├── Customer Service (Level 2+)
└── English Fluency (Level 3+)

Optional Skills:
├── Advanced Certifications (bonus)
├── Spanish Fluency (secondary channel)
└── VIP Customer Experience (specialized)
```

### Bilingual Sales Queue
```
Required Skills:
├── Sales Techniques (Level 3+)
├── Product Knowledge (Level 3+)
├── English Fluency (Level 4+)
├── Spanish Fluency (Level 4+)
└── Customer Service (Level 3+)

Optional Skills:
├── Enterprise Sales (bonus)
├── Account Management (bonus)
└── Negotiation (specialized)
```

---

## Real Flow Scenario: Predictive Routing in Action

```
Customer Calls Support Line
    ↓
System Analyzes: "Technical issue with mobile app"
    ↓
Route Requirements:
├── Skill: Mobile Development (Level 3+)
├── Skill: Customer Service (Level 2+)
├── Language: English
├── Channel: Voice
└── Queue: Technical Support
    ↓
Predictive Engine Evaluates All Available Agents:

Agent 1 (Sarah)
├── Mobile Development: Level 4 ✓
├── Customer Service: Level 4 ✓
├── Current Load: 1 contact
├── Avg Handle Time: 8 mins
├── FCR Rate: 85% ✓ (BEST MATCH)

Agent 2 (James)
├── Mobile Development: Level 2 (Below threshold)
├── Current Load: 2 contacts
├── FCR Rate: 72%

Agent 3 (Maria)
├── Mobile Development: Level 3 ✓
├── Customer Service: Level 3 ✓
├── Current Load: 3 contacts
├── FCR Rate: 78%
    ↓
Route to Agent Sarah (Highest Probability of Resolution)
    ↓
Sarah Answers Call
    ↓
System Logs Interaction for Future Learning
```

---

## Omnichannel Predictive Routing

### Voice Channels
```
Inbound Calls
    ↓
Predictive Routing
    ↓
Skill-based Queue
    ↓
Agent Answer
```

### Digital Channels
```
Chat/Email/Social Arrival
    ↓
Predictive Routing Engine
    ↓
Agent Availability Check (across channels)
    ↓
Route to Agent with Capacity
    ↓
Agent Manages Omnichannel Load
```

### Contact Blending Example
```
Agent Can Handle:
├── 1 Voice Call
├── 2 Chat Conversations
├── 1 Email Thread
└── 1 Social Media Message

Total Capacity: 5 Concurrent Contacts
Current Load: 3 Contacts
Available Slots: 2
```

---

## Real Flow Scenario: Omnichannel Assignment

```
Multiple Contacts in Queue:
├── Inbound Call (Technical)
├── Chat (Billing Question)
└── Email (Complaint)

Predictive Engine Evaluates:
    ↓
Agent 1 Capacity: 2 slots available
├── Skills: Technical, Billing, Customer Service ✓
├── Current: 1 Call + 1 Chat
├── Assignment: Route Email (write capability available)
    ↓
Agent 2 Capacity: 1 slot available
├── Skills: Technical, Billing ✓
├── Current: 1 Chat
├── Assignment: Route Call (available)
    ↓
Queue Assignment Complete
```

---

## Usage Scenarios

| Scenario | Solution | Outcome |
|---|---|---|
| High call volume with variable skill requirements | Enable predictive routing with skill-based queues | Reduced wait times, improved FCR |
| Multiple agent skill levels | Set proficiency thresholds in routing rules | Appropriate complexity matching |
| Omnichannel contact center | Use predictive routing across all channels | Optimized agent utilization |
| International operations | Configure language skills and regional routing | Better customer satisfaction |
| Seasonal staffing fluctuations | Adjust skill assignments dynamically | Maintained service quality |
| VIP customer handling | Create specialized skill for VIP interactions | Enhanced customer experience |

---

## Predictive Routing Configuration Settings

### Queue-Level Settings
```
Queue Configuration: Technical Support

Routing Mode: Predictive Routing ✓

Skill Matching:
├── Required Skills: Product Knowledge, Troubleshooting
├── Proficiency Threshold: Level 3+
├── Language Match: Required
└── Strict Matching: Enabled

Fallback Behavior:
├── If no perfect match: Lower proficiency threshold
├── Escalation path: Senior support queue
└── Timeout: 30 seconds to find match

Load Balancing:
├── Max contacts per agent: 3
├── Considering ACW time: Yes
└── Fair distribution: Enabled
```

---

## Monitoring & Analytics Dashboard

### Key Metrics to Track
| Metric | Target | Purpose |
|---|---|---|
| First Contact Resolution (FCR) | >80% | Measure routing effectiveness |
| Average Handle Time (AHT) | Baseline - 5% | Track efficiency |
| Agent Utilization | 80-90% | Optimize resource use |
| Customer Satisfaction (CSAT) | >85% | Measure outcomes |
| Skill Match Rate | >95% | Verify routing accuracy |
| Queue Abandonment | <5% | Monitor wait times |
| Call Transfer Rate | <10% | Reduce routing errors |

### Real-Time Dashboard View
```
Predictive Routing Performance (Live)

Queue: Technical Support
├── Active Contacts: 24
├── Available Agents: 8
├── Avg Match Score: 4.2/5 ✓
├── Current AHT: 9.2 mins
└── Skill Match %: 96.4%

Top Performing Skills Today:
├── Mobile Development (8 contacts, 87% FCR)
├── Cloud Services (6 contacts, 82% FCR)
└── Hardware Support (4 contacts, 85% FCR)

Agent Performance:
├── Sarah: 4 contacts, 8.1 avg mins, 88% FCR
├── James: 3 contacts, 9.5 avg mins, 79% FCR
└── Maria: 2 contacts, 8.8 avg mins, 84% FCR
```

---

## Best Practices

### Skill Management
- **Keep skills current** - Update agent skills quarterly or after training
- **Avoid over-specialization** - Limit skill count to 8-10 per agent
- **Balance proficiency** - Mix Level 3-4 and Level 2 agents in queues
- **Document requirements** - Clearly define skill needs per queue
- **Regular training** - Invest in skill development to improve proficiency

### Routing Optimization
- **Start with hard rules** - Use required skills and language matching first
- **Monitor thresholds** - Adjust proficiency requirements based on performance
- **Test changes** - Implement rule changes gradually
- **Leverage analytics** - Use data to identify routing improvements
- **Continuous tuning** - Predictive routing improves over time with more data

### Agent Management
- **Clear skill assignments** - Agents must know their assigned skills
- **Growth paths** - Provide training to increase proficiency levels
- **Regular feedback** - Share routing and performance data
- **Incentivize learning** - Reward skill development
- **Load balance fairly** - Ensure equitable work distribution

### Monitoring & Reporting
- **Daily reviews** - Check key metrics and anomalies
- **Weekly analysis** - Identify trends and improvement opportunities
- **Monthly optimization** - Adjust skills and rules as needed
- **Quarterly planning** - Forecast and plan for seasonal changes
- **Annual strategy** - Evaluate overall routing effectiveness

---

## Common Configuration Scenarios

### Scenario 1: Small Team (15 agents, 3 skill levels)
```
Configuration:
├── Premium Edition ✓
├── Workforce Optimization Module ✓
├── Single Queue (Support)
├── 3 Skill Categories (Level 1-4 scaling)
└── Basic Routing Rules

Expected Results:
├── 20-30% improvement in FCR
├── 10-15% reduction in AHT
└── Higher agent satisfaction
```

### Scenario 2: Large Enterprise (200+ agents, multiple queues)
```
Configuration:
├── Premium Edition ✓
├── Workforce Optimization Module ✓
├── 5-8 Queues (Technical, Billing, Sales, etc.)
├── 15+ Skill Categories with proficiency levels
├── Advanced Routing Rules and Overrides
└── Omnichannel Blending

Expected Results:
├── 25-40% improvement in FCR
├── 15-25% reduction in AHT
├── Significant CSAT increase
└── Better resource utilization
```

### Scenario 3: Multilingual Contact Center (5 languages)
```
Configuration:
├── Language Skills for each agent
├── Language Matching in Routing Rules
├── Regional Queue Assignment
├── Skill + Language Combination Routing
└── Overflow to general queue if no match

Expected Results:
├── Improved customer satisfaction
├── Reduced transfers
├── Better first contact resolution
└── International service quality
```

---

## Troubleshooting Guide

| Issue | Cause | Resolution |
|---|---|---|
| Contacts routing to wrong skill level | Proficiency threshold too low | Increase threshold in routing rules |
| Long wait times | Predictive routing disabled for queue | Enable predictive routing in queue config |
| High transfer rates | Skill mismatch in routing | Review and update skill definitions |
| Uneven agent load | Skill imbalance among agents | Provide training to level skill distribution |
| Poor FCR rates | Insufficient skill matching | Add more skill categories to definitions |
| Agents overloaded | Capacity settings too high | Reduce max contacts per agent |
| Low agent utilization | Skill requirements too strict | Relax proficiency thresholds slightly |
| Routing delays | Too many hard rules | Simplify rules and priorities |
| Language mismatch | Language skills not configured | Add language proficiency to agent setup |
| Module not working | Feature not enabled | Verify Premium edition and module purchase |

---

## Real-World Implementation Timeline

### Week 1-2: Assessment & Planning
```
Day 1-2: Kick-off meeting
Day 3-5: Audit current routing and skills
Day 6-10: Design new skill structure
Day 11-14: Plan change management
```

### Week 3-4: Configuration
```
Day 15-18: Create skill definitions
Day 19-22: Configure queues and rules
Day 23-25: Set up monitoring dashboard
Day 26-28: Document configuration
```

### Week 5-6: Testing & Pilot
```
Day 29-32: Conduct routing tests
Day 33-36: Deploy to pilot queue
Day 37-40: Monitor pilot performance
Day 41-42: Gather feedback and optimize
```

### Week 7-8: Full Deployment
```
Day 43-44: Plan rollout schedule
Day 45-52: Deploy to remaining queues
Day 53-56: Monitor and support agents
```

---

## Naming Convention

### Queue Naming with Routing Type
`<Department>_<Function>_<RoutingType>_Queue`

Examples:
- `Support_Technical_PredictiveRouting_Queue`
- `Sales_Enterprise_PredictiveRouting_Queue`
- `Billing_Collections_PredictiveRouting_Queue`

### Skill Naming Convention
`<Category>_<SubCategory>_<Level>`

Examples:
- `Product_MobileApp_Skill`
- `Language_Spanish_Fluency`
- `Service_VIP_Handling`
- `Technical_CloudServices_Certification`

---

## Integration with Other Systems

### Workforce Management (WFM)
```
Predictive Routing ←→ WFM
├── Share agent availability
├── Schedule compliance
├── Forecasting data
└── Staffing adjustments
```

### Quality Management
```
Predictive Routing ←→ Quality Management
├── Interaction recordings
├── Performance metrics
├── Coaching opportunities
└── Training recommendations
```

### Customer Data Platform
```
Predictive Routing ←→ Customer Data
├── Customer history
├── Preferences
├── Previous resolutions
└── VIP status
```

---

## Performance Benchmarks

### Industry Standard Improvements
| Metric | Typical Improvement |
|---|---|
| First Contact Resolution | +15-40% |
| Average Handle Time | -10-25% |
| Customer Satisfaction | +10-25% |
| Agent Utilization | +5-15% |
| Queue Abandonment | -20-50% |
| Cost Per Contact | -10-20% |

### Ramp-Up Timeline
```
Day 1-30: Learning phase, marginal improvement
Month 2-3: System learns patterns, 10-15% improvement
Month 4-6: Optimized configuration, 25-35% improvement
Month 6+: Mature state, 30-40% improvement
```

---

## Predictive Routing vs. Traditional Routing

| Feature | Predictive Routing | Traditional Routing |
|---|---|---|
| Skill Matching | AI-optimized | Rules-based |
| Agent Selection | Predictive | Sequential |
| Learning | Continuous | None |
| Complexity | High | Low |
| Setup Time | Medium | Low |
| Optimization | Automatic | Manual |
| FCR Improvement | 25-40% | Baseline |
| Cost | Higher | Lower |

---

## Interview Cheat Sheet

| Question | Answer |
|---|---|
| What is Predictive Routing? | AI-powered routing system that optimizes agent assignment using ML |
| What are the requirements? | Premium edition + Workforce Optimization module |
| How does it select agents? | Analyzes skills, availability, performance, and predicts best match |
| What factors does it consider? | Skills, proficiency, availability, workload, language, history |
| Can you use it with omnichannel? | Yes, works with voice, chat, email, and messaging |
| How is it different from skill-based routing? | Uses ML to predict best match vs. just checking skills |
| Where do you configure it? | Admin → Contact Center → Routing |
| What's the expected improvement? | FCR +25-40%, AHT -10-25%, CSAT +10-25% |
| How long does setup take? | 4-8 weeks depending on complexity |
| What are critical success factors? | Accurate skills data, proper proficiency levels, continuous monitoring |
| How do you monitor performance? | Daily dashboard review, weekly analytics, monthly optimization |
| What if no perfect agent match exists? | System queues contact with fallback routing rules |
| Can you override predictive routing? | Yes, hard rules take precedence (availability, required skills) |
| How does machine learning help? | Learns from past outcomes to improve future routing |
| What's the ROI timeline? | 2-3 months to see significant improvements |

---

## Key Takeaways

- **AI-Powered Optimization** - Predictive routing uses machine learning to match contacts to best-fit agents
- **Skill-Based Foundation** - Requires well-defined skills and proficiency levels
- **Omnichannel Capable** - Works across voice, chat, email, and messaging channels
- **Continuous Learning** - System improves routing decisions over time
- **Premium Feature** - Requires Premium edition and Workforce Optimization module
- **Significant ROI** - Typical improvements: FCR +25-40%, AHT -10-25%
- **Real-Time Optimization** - Routes contacts dynamically based on current system state
- **Fallback Rules Matter** - Hard rules ensure quality even without perfect matches
- **Change Management Critical** - Proper implementation and monitoring are essential
- **Ongoing Monitoring Required** - Daily reviews and monthly tuning maximize benefits

---

## Migration Path from Traditional Routing

### Phase 1: Preparation (Weeks 1-2)
```
├── Audit current routing rules
├── Document all skills currently used
├── Identify skill gaps
└── Plan queue restructuring
```

### Phase 2: Setup (Weeks 3-4)
```
├── Create comprehensive skill definitions
├── Assign skills and proficiency to agents
├── Configure predictive routing queues
└── Establish monitoring dashboards
```

### Phase 3: Pilot (Weeks 5-6)
```
├── Enable on low-risk queue
├── Monitor closely for issues
├── Gather team feedback
└── Optimize configuration
```

### Phase 4: Rollout (Weeks 7-8)
```
├── Disable traditional routing rules
├── Enable predictive on remaining queues
├── Support agents through transition
└── Celebrate early wins
```

---

## Additional Resources

### Official Documentation Links
- Genesys Cloud Routing Guide: https://help.genesys.com/genesyscloud/current/en-us/Routing.html
- Predictive Routing Setup: https://help.genesys.com/genesyscloud/current/en-us/PredictiveRouting.html
- Workforce Optimization: https://help.genesys.com/genesyscloud/current/en-us/WFO.html

### Support Contacts
- Genesys Sales: sales@genesys.com
- Genesys Support: https://support.genesys.com
- Community Forums: https://community.genesys.com

---

## Document Version Info
**Last Updated:** March 2026  
**Source:** Genesys PureCloud Official Documentation  
**Version:** 1.0