Forecasting

Genesys WFM Forecasting Documentation 
 Study Notes 
 
 
 
 Topic 
 Description 
 
 
 
 
 Master Forecast 
 Published forecast scenario used for scheduling 
 
 
 Automatic Best Method 
 AI analyzes 10+ algorithms to select best forecast 
 
 
 Forecasting Methods 
 ABM, WHI, HDI, Import Forecast 
 
 
 Main Forecast 
 Continuous daily calculation from latest data 
 
 
 Accuracy 
 ABM: 85-92% vs Traditional: 70-80% 
 
 
 Planning Groups 
 Organize workload by media type and route 
 
 
 Service Goals 
 Define SL, ASA, abandon rate targets 
 
 
 Recalculation 
 Main forecast recalculates nightly with new data 
 
 
 
 
 Navigation 
 Admin → Workforce Management → Forecasting
OR
Menu → Workforce Management → Forecasting → Forecasts 
 
 Forecasting Overview 
 Forecasting is the process of predicting future contact volume and average handle time to determine required staffing levels. WFM uses forecasts to create optimal schedules that balance service level goals with operational efficiency. 
 Forecasts form the foundation of workforce planning. Accurate forecasts drive better schedules, which drive better adherence, which drives better service levels. The forecasting process involves analyzing historical data, selecting appropriate forecasting methods, validating scenarios, and publishing the best scenario to become the Master Forecast. 
 Forecasting Objectives 
 
 Volume Prediction - Predict number of interactions (offered) 
 AHT Prediction - Predict average handle time per interaction 
 Staffing Calculation - Determine agents needed to meet service goals 
 Scenario Planning - Evaluate multiple forecasting approaches 
 Continuous Improvement - Refine based on actual vs forecast variance 
 Capacity Planning - Support long-range hiring decisions 
 
 Forecasting Scope 
 
 Business Unit Level - Forecasts created for entire BU (all MUs) 
 Time Granularity - Hourly and daily forecasts 
 Planning Period - 26 weeks forward (can search up to 104 weeks) 
 Historical Data - Requires 90+ days of data for accuracy 
 Multiple Scenarios - Create and compare multiple approaches 
 Master Forecast - One published forecast per BU at a time 
 
 
 Forecasting Methods 
 1. Automatic Best Method (ABM) 
 Automatic Best Method is the AI-powered forecasting approach that analyzes historical interaction data and automatically selects the most accurate forecasting algorithm from 10+ methods. 
 How ABM Works: 
 Historical Data Input
 ↓
Analyze 10+ Forecasting Algorithms:
├─ Moving Average
├─ Exponential Smoothing
├─ Trend Analysis
├─ Seasonal Decomposition
├─ Regression Models
├─ Time Series Analysis
└─ ... 4+ additional models
 ↓
Evaluate Each Against Historical Data
├─ Calculate accuracy metrics
├─ Test fit quality
├─ Assess seasonal patterns
└─ Validate trend capture
 ↓
Select Best Performing Model
├─ Lowest error rate
├─ Best seasonal fit
├─ Most stable predictions
└─ Confidence validation
 ↓
Generate Forecast
├─ Volumes (Offered)
├─ AHT (Average Handle Time)
└─ Staffing Requirements
 
 ABM Characteristics: 
 
 Requires minimum 90 days historical data 
 Automatically evaluates 10+ algorithms 
 Selects best fit without manual intervention 
 Accuracy: 85-92% (vs traditional 70-80%) 
 Recalculates nightly with new data 
 Supports all media types 
 Cloud-based processing 
 
 When to Use ABM: 
 
 ✓ Mature contact center (90+ days data) 
 ✓ Accurate historical data available 
 ✓ Want AI-driven optimization 
 ✓ Looking for best accuracy 
 ✓ Limited forecasting expertise 
 
 ABM Limitations: 
 
 ✗ Requires 90+ days data minimum 
 ✗ Cannot account for business events manually 
 ✗ May not work with highly volatile data 
 ✗ Limited control over algorithm selection 
 
 2. Weighted Historical Index (WHI) 
 Weighted Historical Index allows forecasters to assign importance to specific historical periods, enabling forecasts that reflect anticipated trends or business changes. 
 How WHI Works: 
 Select Historical Periods
 ↓
Assign Weights to Periods:
├─ Recent period: 40% weight (higher importance)
├─ Same season last year: 35% weight
├─ 2 years ago: 15% weight
└─ Older data: 10% weight
 ↓
Calculate Weighted Average
├─ Multiply volumes by weights
├─ Multiply AHT by weights
└─ Generate forecast
 ↓
Manual Adjustments
├─ Add/subtract for known events
├─ Adjust for staffing changes
└─ Account for market changes
 ↓
Forecast Output
 
 WHI Characteristics: 
 
 Manual weight assignment 
 Incorporates business judgment 
 Good for known changes 
 Moderate accuracy (75-85%) 
 Requires forecasting expertise 
 Supports business events 
 
 When to Use WHI: 
 
 ✓ Known business changes upcoming 
 ✓ New product launch planned 
 ✓ Merger/acquisition integration 
 ✓ Staffing changes anticipated 
 ✓ Want forecaster control 
 
 WHI Limitations: 
 
 ✗ Requires manual configuration 
 ✗ Subject to forecaster bias 
 ✗ More time-consuming setup 
 ✗ Less accurate than ABM typically 
 
 3. Historical Data Import (HDI) 
 Historical Data Import enables importing external historical data via CSV files, useful for organizations lacking internal historical data or integrating data from legacy systems. 
 How HDI Works: 
 Prepare Historical Data CSV:
├─ Date, Time, Interactions Offered
├─ Date, Time, Average Handle Time
└─ Format per Genesys specifications

Upload CSV File
 ↓
Data Validation
├─ Check date formats
├─ Validate interaction counts
├─ Verify AHT values
└─ Check for gaps

Map to Planning Groups
├─ Assign data to queues/routes
├─ Set media types
└─ Configure mapping rules

Import and Store
├─ Historical data added to system
├─ Available for forecasting
└─ Retained for 90+ days

Create Forecast
├─ ABM or WHI using imported data
├─ Generate volumes and AHT
└─ Publish to master forecast
 
 HDI Characteristics: 
 
 CSV-based import 
 Supports external data sources 
 Integrates legacy system data 
 Enables quick setup for new centers 
 Maintains data history 
 Works with ABM/WHI methods 
 
 When to Use HDI: 
 
 ✓ New contact center (no history) 
 ✓ Migrating from legacy system 
 ✓ Acquiring another company 
 ✓ Need to supplement internal data 
 ✓ Have external forecasting data 
 
 HDI Limitations: 
 
 ✗ Requires proper CSV formatting 
 ✗ Data validation needed 
 ✗ Manual upload process 
 ✗ Can't import future predictions 
 
 4. Import Forecast 
 Import Forecast supports uploading externally generated forecasts into Genesys Cloud, integrating with existing forecasting systems or third-party tools. 
 How Import Forecast Works: 
 External Forecast Generation:
├─ Create in Excel/third-party tool
├─ Calculate volumes and AHT
├─ Format per specifications
└─ Validate accuracy

Export as CSV/XML
 ↓
Upload to Genesys WFM
├─ Select planning group
├─ Map forecast period
└─ Validate format

System Processing:
├─ Parse forecast data
├─ Validate ranges
├─ Check for completeness
└─ Store in system

Publish to Master Forecast
├─ Available for scheduling
├─ Used for staffing calculations
└─ Drives schedule generation
 
 Import Forecast Characteristics: 
 
 Forecast generated externally 
 Upload pre-calculated volumes/AHT 
 Bypasses ABM/WHI 
 Useful for specialized models 
 One-time or periodic imports 
 No recalculation 
 
 When to Use Import Forecast: 
 
 ✓ Have specialized forecasting tool 
 ✓ Third-party forecast provider 
 ✓ Complex custom models 
 ✓ Forecasting done elsewhere 
 ✓ Minimal WFM forecasting expertise 
 
 Import Forecast Limitations: 
 
 ✗ No automatic updates 
 ✗ Must re-import when data changes 
 ✗ No integration with real-time data 
 ✗ Less adaptive to changes 
 
 
 Main Forecast 
 The Main Forecast is a special forecast calculated continuously (typically nightly) based on all available historical data. It provides the baseline forecast that can be used immediately or modified through scenarios. 
 Main Forecast Characteristics: 
 
 Continuous Calculation - Recalculates every night 
 All Historical Data - Uses complete data history 
 Automatic - No manual intervention required 
 Read-Only - Cannot be directly edited 
 Baseline Reference - Starting point for scenarios 
 Always Available - Immediately accessible 
 Best Current Data - Uses latest interactions 
 
 Main Forecast Flow: 
 Day 1: Historical Data (30 days)
 ↓ Nightly Calculation
 ↓ Main Forecast Published
 ↓
Day 2: Historical Data (30 days) + Day 1 New Data
 ↓ Nightly Calculation
 ↓ Main Forecast Updated
 ↓
Ongoing: Continuous refinement with new daily data
 
 
 Planning Groups 
 Planning Groups organize workloads by specific media types and route paths, enabling targeted forecasting and scheduling. 
 Planning Group Components: 
 Planning Group: Support - Inbound Voice
├─ Media Type: Voice (inbound)
├─ Queue/Route: Support_Queue_001
├─ Skill Required: Support_Skill_Level_2+
├─ Service Goal: 80% SL, 20 sec ASA, 5% abandon
├─ Staffing Group: Support_Agents
└─ Forecast Data: Volume + AHT

Planning Group: Sales - Outbound Calls
├─ Media Type: Voice (outbound)
├─ Campaign: Q1_Spring_Campaign
├─ Skill Required: Sales_Skill_Level_1+
├─ Service Goal: Contact 50% of list, 5 min calls
├─ Staffing Group: Sales_Agents
└─ Forecast Data: Dialing ratio + AHT

Planning Group: Support - Email
├─ Media Type: Email
├─ Queue: Support_Email_Queue
├─ Response Time Goal: 4 hours
├─ Staffing Group: Support_Agents
└─ Forecast Data: Volume + AHT

Planning Group: Chat Support
├─ Media Type: Chat
├─ Route: Support_Chat_Route
├─ Concurrency: 4-5 chats per agent
├─ Response Time: Immediate
└─ Forecast Data: Offered + AHT
 
 Planning Group Creation: 
 
 Navigate to Admin → Workforce Management → Planning Groups 
 Click Create Planning Group 
 Configure:
 
 Name - Unique identifier 
 Business Unit - Parent BU 
 Media Type - Voice, Email, Chat, Callback, Messaging, Workitems 
 Queue/Route - Associated queue or route 
 Service Goal - Target metrics 
 Staffing Model - Agents to use 
 
 
 Save and activate 
 
 Planning Group Best Practices: 
 
 Clarity - Clear names reflecting function 
 Consolidation - Group similar work together 
 Skill Alignment - Agents have required skills 
 Service Goals - Match business objectives 
 Regular Review - Update as operations change 
 Documentation - Maintain planning group mapping 
 
 
 Service Goals 
 Service Goals define the performance targets for a planning group: Service Level, Average Speed to Answer, and Abandonment Rate. 
 Service Goal Components: 
 Service Goal Template: Premium Support
├─ Service Level Goal: 80%
│ └─ Definition: 80% of calls answered within 20 seconds
├─ Average Speed to Answer: 20 seconds
│ └─ Definition: Average answer time across all calls
├─ Abandon Rate: 5%
│ └─ Definition: Max 5% of offered calls abandoned
├─ Media Type: Voice
├─ Time Intervals: Hourly
└─ Period: Weekly

Service Goal Template: Standard Support
├─ Service Level Goal: 75%
├─ Average Speed to Answer: 30 seconds
├─ Abandon Rate: 8%
└─ More lenient targets for lower-volume periods

Service Goal Template: Email Support
├─ Service Level Goal: 95% within 4 hours
├─ Average Speed to Answer: 2 hours (median response)
├─ Abandon Rate: 0% (not applicable)
└─ Media Type: Email
 
 Service Goal Best Practices: 
 
 Realistic Targets - Achieve 80-85% of time 
 Business Aligned - Match customer expectations 
 Media-Specific - Different for voice vs email 
 Documented - Communicate to teams 
 Regular Review - Adjust based on performance 
 Incremental Improvement - Tighten gradually 
 
 
 Forecasting Process 
 Step 1: Data Preparation 
 
 Ensure 90+ days of historical data available 
 Validate data accuracy 
 Check for gaps or anomalies 
 Clean outliers if necessary 
 Confirm queue/route mappings 
 
 Step 2: Create Scenario 
 
 Navigate to Forecasts → Scenarios 
 Click New Scenario 
 Configure:
 
 Name - e.g., "Q2_2026_Base_Forecast" 
 Period Start - Beginning of forecast 
 Period End - 26 weeks forward 
 Planning Groups - Select which to include 
 
 
 Create scenario 
 
 Step 3: Build Volumes 
 
 Open scenario 
 Click Build Volumes 
 Select forecasting method:
 
 ABM - Recommended for accuracy 
 WHI - For known business changes 
 Template - Copy from similar period 
 
 
 Configure method-specific settings 
 Generate volumes 
 
 Step 4: Build AHT 
 
 Open scenario volumes 
 Click Build AHT 
 Select method (typically same as volumes) 
 Configure AHT-specific settings 
 Generate AHT 
 
 Step 5: Review & Validate 
 
 Open Scenario Volumes view 
 Review volume trends:
 
 ✓ Match business expectations 
 ✓ Seasonal patterns visible 
 ✓ Growth/decline appropriate 
 ✓ No obvious anomalies 
 
 
 Open Scenario Staffing view 
 Review staffing requirements:
 
 ✓ Realistic agent counts 
 ✓ Service level achievable 
 ✓ Aligned with budget 
 ✓ Growth manageable 
 
 
 
 Step 6: Compare Scenarios 
 
 Create multiple scenarios if desired 
 Compare side-by-side:
 
 Volume projections 
 Staffing requirements 
 Cost implications 
 Service level achievement 
 
 
 Select best scenario 
 
 Step 7: Publish to Master Forecast 
 
 Open best scenario 
 Click Publish to Master Forecast 
 Confirm publication 
 Master Forecast now available for scheduling 
 
 Step 8: Monitor & Adjust 
 
 Track actual vs forecast weekly 
 Calculate variance:
 
 Volume Variance = (Actual - Forecast) / Forecast 
 AHT Variance = (Actual - Forecast) / Forecast 
 
 
 Adjust future forecasts based on variance 
 Recalculate Main Forecast 
 
 
 Forecasting Accuracy 
 Measuring Accuracy 
 Volume Accuracy: 
 Forecast Accuracy = 1 - |Actual - Forecast| / Actual

Example:
Forecasted Offered: 1,000 calls
Actual Offered: 980 calls
Variance: |980 - 1,000| / 1,000 = 2%
Accuracy: 98%
 
 AHT Accuracy: 
 Forecasted AHT: 420 seconds
Actual AHT: 410 seconds
Variance: |410 - 420| / 420 = 2.4%
Accuracy: 97.6%
 
 Overall Forecast Accuracy: 
 
 Excellent - 95%+ accuracy 
 Good - 90-95% accuracy 
 Acceptable - 85-90% accuracy 
 Poor - <85% accuracy 
 
 Factors Affecting Accuracy 
 Positive Factors: 
 
 ✓ Abundant historical data (90+ days) 
 ✓ Consistent seasonal patterns 
 ✓ Stable agent population 
 ✓ No major business changes 
 ✓ Accurate data collection 
 ✓ Use of ABM method 
 
 Negative Factors: 
 
 ✗ Insufficient historical data (<30 days) 
 ✗ Volatile contact volumes 
 ✗ Seasonal anomalies 
 ✗ Major staffing changes 
 ✗ Business discontinuities 
 ✗ Data quality issues 
 
 Improving Forecast Accuracy 
 
 
 Data Quality 
 
 Validate interaction data 
 Remove duplicate entries 
 Correct time stamps 
 Clean outliers appropriately 
 
 
 
 Time Frame Selection 
 
 Use 90+ days of data 
 Include full seasonal cycle 
 Exclude anomalous periods 
 Weight recent data higher 
 
 
 
 Method Selection 
 
 Use ABM for stable patterns 
 Use WHI for known changes 
 Test multiple methods 
 Compare results 
 
 
 
 Ongoing Monitoring 
 
 Track actual vs forecast weekly 
 Identify variance sources 
 Adjust future forecasts 
 Document lessons learned 
 
 
 
 Business Communication 
 
 Inform of known changes 
 Get campaign dates in advance 
 Understand staffing constraints 
 Align on service goals 
 
 
 
 
 Real-World Examples 
 Example 1: New Contact Center (Using HDI) 
 Scenario: New financial services contact center
Problem: No historical data in Genesys
Solution: Historical Data Import

Process:
1. Obtain 6 months of data from legacy system
2. Format as CSV (Date, Time, Volume, AHT)
3. Upload to WFM via Historical Data Import
4. Validate imported data (1,000+ calls/day confirmed)
5. Create ABM forecast using imported data
6. Generate 26-week forecast with 85% confidence
7. Publish to Master Forecast
8. Create schedules based on forecast
9. Begin tracking actual vs forecast
10. Refine with real Genesys data over time

Result: Forecast accuracy 82% week 1, improving to 90% by week 12
 
 Example 2: Seasonal Business (Using WHI) 
 Scenario: Retail customer service (high holiday season)
Problem: Regular ABM doesn't account for expected surge
Solution: Weighted Historical Index

Process:
1. Run ABM to get baseline forecast
2. Weight recent 4 weeks: 50%
3. Weight same season last year: 30%
4. Weight 2 years ago: 20%
5. Manual adjustment: +25% for new catalog
6. Manual adjustment: +10% for holiday promotions
7. Result: 15,000 calls/day (vs ABM baseline 12,000)
8. Schedule accordingly with additional flex agents
9. Publish to Master Forecast
10. Monitor first week for variance

Result: Forecast accuracy 88% during peak season
Alternative: Would have been 65% with unmodified ABM
 
 Example 3: Business Event (Using Import Forecast) 
 Scenario: Product launch with external forecast
Problem: Marketing has already forecasted demand impact
Solution: Import Forecast from external source

Process:
1. Marketing provides forecast:
 - Week 1: +30% volume increase
 - Week 2: +50% volume increase
 - Week 3: +40% volume increase
 - Week 4-8: Declining to normal
2. Obtain volumes from marketing
3. Format as CSV per Genesys spec
4. Upload via Import Forecast
5. Map to Support planning group
6. Validate import completeness
7. Publish to Master Forecast
8. Create high-staffing schedule for weeks 1-3
9. Monitor actual vs marketing forecast
10. Adjust staffing based on real results

Result: Forecast accuracy 92% (marketing expertise applied)
Cost: Hired 50 temporary agents, all utilized during surge
 
 
 Best Practices 
 Forecasting Process 
 
 Establish Baseline - Start with ABM for consistency 
 Regular Reviews - Monthly variance analysis 
 Scenario Planning - Test multiple approaches 
 Documentation - Record assumptions and decisions 
 Communication - Share forecasts with operations 
 Validation - Compare to business expectations 
 
 Data Management 
 
 Data Quality - Daily validation and cleanup 
 Retention - Keep 90+ days for pattern recognition 
 Integrity - Prevent manual edits without documentation 
 Backup - Maintain forecast history 
 Audit Trail - Track who changed what when 
 
 Continuous Improvement 
 
 Weekly Tracking - Actual vs forecast variance 
 Root Cause Analysis - Understand variances 
 Method Adjustment - Change weights/methods as needed 
 Feedback Loop - Share insights with business 
 Seasonal Adjustment - Update for known patterns 
 
 
 Interview Cheat Sheet 
 
 
 
 Question 
 Answer 
 
 
 
 
 What's ABM? 
 Automatic Best Method - AI selects best algorithm from 10+ 
 
 
 ABM accuracy? 
 85-92% (vs traditional 70-80%) 
 
 
 ABM data requirement? 
 Minimum 90 days historical data 
 
 
 When use ABM? 
 Mature center, good data, want best accuracy 
 
 
 When use WHI? 
 Known business changes, want forecaster control 
 
 
 When use HDI? 
 New center, migrating systems, external data 
 
 
 When use Import? 
 External forecast available, specialized models 
 
 
 What's Main Forecast? 
 Automatically calculated nightly using all data 
 
 
 What's planning group? 
 Organizes work by media type and route 
 
 
 What's service goal? 
 Targets for SL, ASA, abandon rate 
 
 
 Where forecasts created? 
 Business Unit level 
 
 
 How far forward? 
 26 weeks (can search up to 104) 
 
 
 How often recalculate? 
 Main forecast nightly, scenarios on demand 
 
 
 Forecast impacts? 
 Drives scheduling, staffing, service level 
 
 
 How measure accuracy? 
 Compare actual vs forecast volume and AHT 
 
 
 
 
 Key Takeaways 
 
 AI-Powered - ABM automatically selects best method from 10+ algorithms 
 Accuracy - ABM achieves 85-92% vs traditional 70-80% 
 Flexibility - Four methods (ABM, WHI, HDI, Import) for different scenarios 
 Continuous - Main Forecast recalculates nightly with new data 
 Planning Groups - Organize by media type and route for precise forecasting 
 Service Goals - Define targets for SL, ASA, abandon rate 
 Data-Driven - Requires 90+ days of data for accuracy 
 Validation - Regular monitoring of actual vs forecast variance 
 Business Events - WHI and Import methods support known changes 
 Foundation - Forecast quality directly impacts schedule quality 
 
 
 Additional Resources 
 Official Documentation 
 
 Forecasting Overview: help.genesys.cloud/articles/forecasting-overview/ 
 Automatic Best Method: help.genesys.cloud/articles/automatic-best-method/ 
 Planning Groups: help.genesys.cloud/articles/planning-groups-overview/ 
 Service Goals: help.genesys.cloud/articles/service-goals-overview/ 
 
 Support & Training 
 
 Genesys University: genesys.com/training 
 Community Forums: https://community.genesys.com 
 Technical Support: https://support.genesys.com 
 
 
 Document Version Info 
 Last Updated: March 2026 
 Source: Genesys WFM Official Documentation 
 Validated: Current with January-March 2026 releases 
 Version: 1.0