Average Handle Time
Average handle time (AHT) is a workforce management metric that measures the mean duration of a customer interaction from initiation to completion, including all associated wrap-up work. It is one of the most fundamental inputs to workforce management staffing calculations — alongside contact volume, AHT determines the total workload (traffic intensity in Erlangs) that must be covered by scheduled agents.
AHT serves dual roles in contact center operations: as a planning metric (how long agents need per contact, driving staffing requirements) and as an efficiency metric (how productively agents handle interactions). These roles can conflict, creating one of the most persistent tensions in WFM practice.
Definition and Components
AHT is the sum of three components:
| Component | Definition | Typical Proportion |
|---|---|---|
| Talk Time | Active conversation between agent and customer | 60-75% of AHT |
| Hold Time | Customer placed on hold while agent researches or consults | 5-15% of AHT |
| After-Call Work (ACW) | Post-interaction tasks: notes, disposition codes, follow-up scheduling, system updates | 15-30% of AHT |
The formula is applied as an average across all contacts in a given period:
where n is the number of contacts handled.
Transfer Time
Some organizations include transfer time as a fourth component — the time spent routing a call to another agent or department. Whether transfer time is included depends on how the ACD system attributes the contact: if the transferred interaction creates a new contact record, each leg has its own AHT; if it remains a single record, transfer time is embedded.
Role in Workforce Management
Staffing Calculations
AHT is a critical input to the Erlang C staffing model. The workload (traffic intensity) for a given interval is:
Small changes in AHT have significant staffing implications. For a center handling 1,000 calls per day at an AHT of 360 seconds (6 minutes), a 10% increase to 396 seconds requires approximately 10 additional FTEs — a substantial cost impact.
Forecasting
AHT forecasting is as important as volume forecasting but receives less attention in many organizations. AHT varies by:
- Time of day: Morning AHTs often differ from afternoon (agent fatigue, customer complexity patterns)
- Day of week: Monday volumes skew toward complex accumulated issues; Friday AHT may drop as agents compress calls
- Contact type: Sales calls, billing inquiries, and technical support have materially different AHTs
- Agent tenure: New agents typically handle calls 30-50% longer than experienced agents
- Channel: Voice, chat, and email have fundamentally different handle time characteristics
Forecasting AHT separately from volume allows WFM teams to detect and respond to trends that would be invisible in a single workload forecast.
AHT by Channel
| Channel | Typical AHT Range | Measurement Notes |
|---|---|---|
| Inbound voice | 4-8 minutes | Most standardized measurement; includes talk + hold + ACW |
| Outbound voice | 3-10 minutes | Highly variable by campaign; includes dial time in some systems |
| Chat | 8-15 minutes | Agent handles multiple concurrent sessions; AHT per session vs. agent time diverge |
| 5-15 minutes | Measured as composition time; does not include queue time | |
| Social media | 3-10 minutes | Often excludes research time between responses |
| Back office | Minutes to hours | Knowledge work AHT varies by case complexity; often uses median rather than mean |
Chat Concurrency and AHT
In chat environments, the distinction between contact AHT and agent AHT is critical. An agent handling three concurrent chats may spend 12 minutes on each chat (contact AHT = 12 minutes), but the total agent time consumed is closer to 15-18 minutes (agent AHT), not 36 minutes. WFM systems must account for concurrency when converting AHT to staffing requirements.
Criticisms and Tensions
AHT as an Efficiency Target
Using AHT as a performance target for individual agents is one of the most debated practices in contact center management:
Arguments for AHT targets:
- Identifies agents who may need coaching on system navigation or call control
- Prevents unnecessary call elongation
- Directly impacts staffing costs
Arguments against AHT targets:
- Agents rush calls to hit targets, reducing first contact resolution
- Customers with complex needs receive inadequate service
- Agents game metrics by transferring calls or shortening ACW at the cost of data quality
- The metric incentivizes speed over quality, undermining customer experience
Research in operations management has shown that aggressive AHT reduction often increases repeat contacts, producing higher total cost despite lower per-contact cost.[1]
The Mean vs. Distribution Problem
AHT is a mean — and contact handle time distributions are typically right-skewed (long tail of complex contacts). This creates problems:
- A small number of long calls can significantly inflate AHT, distorting staffing calculations
- Trimmed means or medians may better represent the "typical" contact but understate workload
- Percentile analysis (e.g., P90 handle time) is more informative for understanding variability
Some organizations forecast both the mean and the distribution shape, using probabilistic methods to capture the uncertainty inherent in AHT variation.
AHT Reduction Strategies
Legitimate AHT reduction focuses on removing friction rather than pressuring agents:
- Knowledge base improvements: Faster access to accurate information reduces research time
- Desktop integration: Unified agent desktops eliminate system-switching between CRM, billing, and knowledge tools
- Automated after-call work: AI-generated call summaries and auto-disposition reduce ACW
- Call flow optimization: Better IVR design routes contacts to the right agent, reducing transfers and repeat explanations
- Agent training: Improved product knowledge and call control skills
- Process simplification: Reducing the number of steps required to resolve common issues
These approaches reduce AHT while maintaining or improving quality — unlike target pressure, which trades quality for speed.
AHT and AI
Artificial intelligence is transforming both the measurement and management of AHT:
- Real-time agent assist: AI suggests responses and surfaces knowledge during the interaction, reducing research time
- Automated summarization: Post-call summaries generated by AI can reduce ACW by 30-60%
- Predictive routing: Matching contact complexity to agent skill reduces handle time through better allocation
- [[Agentic AI Workforce Planning|AI agents]]: Fully automated interactions have zero traditional AHT; the metric shifts to resolution time and accuracy
As AI agents handle an increasing share of contacts, AHT becomes less relevant as a universal metric and more specific to human-handled interactions.
Maturity Model Position
AHT understanding and management evolves across maturity levels:
- Level 1 (Reactive): AHT tracked as a single daily average. Used primarily as an agent performance target.
- Level 2 (Foundational): AHT tracked by interval, contact type, and agent group. Separated from agent performance evaluation.
- Level 3 (Integrated): AHT forecasted independently from volume. Distribution analysis supplements mean tracking.
- Level 4 (Optimized): AHT reduction driven by automation and process optimization rather than targets. Agent-level AHT used for coaching, not scoring.
- Level 5 (Adaptive): AHT distinguished between human and AI contacts. Total resolution time (including automated steps) replaces AHT as the primary metric.
See Also
- Workforce Management — Overview of the WFM discipline
- Service Level — Primary metric that AHT staffing calculations support
- Occupancy — Related efficiency metric
- Erlang C — Staffing model that uses AHT as a key input
- Shrinkage — Non-productive time excluded from AHT
- Forecasting Methods — AHT forecasting as a WFM practice
- Forecast Accuracy Metrics — Measuring AHT forecast quality
- First Contact Resolution — Quality metric that AHT targets can undermine
- Quality Management — The quality-efficiency tension
References
- ↑ Aksin, Zeynep; Armony, Mor; Mehrotra, Vijay (2007). "The Modern Call Center: A Multi-Disciplinary Perspective on Operations Management Research." Production and Operations Management, 16(6), 665-688.
