Average Handle Time in Call Centers — How to Reduce It Without Destroying Quality
Average handle time (AHT) is the total time an agent spends on a customer interaction — from the moment the call connects to the moment the agent finishes their after-call work and becomes available for the next contact. It is one of the most watched metrics in call center operations because it directly affects staffing requirements, cost per call, and how many customers can be served per hour.
It is also one of the most dangerous metrics to optimize in isolation. A call center that pressures agents to reduce AHT without tracking first contact resolution (FCR) will produce shorter calls that do not solve the customer's problem — generating callbacks that increase total handle time, reduce customer satisfaction, and cost more than the original longer call would have.
The goal is not the lowest possible AHT. The goal is the lowest AHT that maintains or improves FCR. This post covers how to analyze AHT by component, identify where time is actually being wasted, and apply fixes that reduce handle time without creating downstream problems.
AHT components
AHT is not one number — it is three:
AHT = Talk Time + Hold Time + After-Call Work (ACW)
Each component has different causes and different fixes. An operation with a 7-minute AHT where 5 minutes is talk time, 30 seconds is hold time, and 90 seconds is ACW has a fundamentally different problem than one where 3 minutes is talk time, 2.5 minutes is hold time, and 90 seconds is ACW. The first operation may be fine — the calls are just complex. The second has a hold time problem that is inflating AHT.
| Component | Typical range | What it measures |
|---|---|---|
| Talk time | 3–6 minutes | Time the agent and customer are in conversation |
| Hold time | 15–60 seconds | Time the customer is on hold while the agent looks up information, consults a supervisor, or navigates a system |
| After-call work (ACW) | 30–90 seconds | Time the agent spends on documentation, disposition codes, or follow-up tasks after the call ends |
| Total AHT | 4–8 minutes | Sum of all three (varies significantly by call type) |
Before trying to reduce AHT, break it into these components and determine which one is the actual problem.
Diagnosing where time is lost
High talk time
Talk time is the conversation itself. When it is high, the cause is either the calls are genuinely complex (and the talk time is appropriate) or agents are inefficient in how they conduct the conversation.
| Cause | How to identify | Fix |
|---|---|---|
| Call complexity | AHT is high for a specific call type but FCR is also high — agents are taking the time needed to resolve complex issues | Not a problem to fix. Set AHT targets by call type rather than using a single target across all call types |
| Poor call control | Agents let customers lead the conversation, go off-topic, or repeat the same information multiple times | Coaching on call control techniques — acknowledging the customer's concern, then directing the conversation toward resolution |
| Inadequate product/process knowledge | Agents spend time during the call reading documentation or figuring out how to handle the issue | Training on the top 10–15 call types that account for 80% of volume. Agents should know the resolution path before the call, not discover it during |
| Unnecessary verification steps | Agents ask for information the system already has, or verify the same details multiple times during a transfer | Streamline verification to the minimum required by policy. If the IVR collected the account number, the agent should not ask for it again |
| Customer venting | Agent does not know how to transition from listening to the complaint to resolving the issue | Train on the acknowledge-bridge-resolve pattern: acknowledge the frustration, bridge to the resolution ("Let me fix that for you right now"), then resolve |
High hold time
Hold time is almost always a process or system problem, not an agent skill problem. When an agent puts a customer on hold, it is because they need something they do not have — information, system access, supervisor approval, or knowledge of the resolution.
| Cause | How to identify | Fix |
|---|---|---|
| System navigation | Agents toggle between multiple screens or applications to find information | Reduce the number of systems agents use. If the CRM, knowledge base, and order system are three separate applications, agents spend hold time switching between them |
| Looking up policies or procedures | Agents put customers on hold to read the knowledge base or ask a neighbor | Improve the knowledge base — if agents cannot find the answer within 15 seconds, the KB structure is the problem, not the agent |
| Supervisor escalation | Agents must get supervisor approval for common actions (credits, exceptions, escalations) | Increase agent authority for routine decisions. If 80% of supervisor approvals are granted, the approval requirement is adding hold time without adding value |
| Warm transfers | Agent holds while waiting for the transfer destination to answer | Track transfer hold time separately. If the receiving team takes 3+ minutes to answer, that team's staffing is the problem |
| System performance | Slow application load times force agents to wait | Monitor system response times. If the CRM takes 8 seconds to load a customer record, every call includes that delay |
Hold time reduction has the highest ROI because it reduces AHT without affecting conversation quality. An agent who spends 30 fewer seconds on hold per call handles 1 additional call per hour — at zero cost to quality.
High after-call work (ACW)
ACW is the time between when the call ends and when the agent is available for the next call. It covers documentation, disposition coding, and any follow-up actions.
| Cause | How to identify | Fix |
|---|---|---|
| Excessive documentation requirements | Agents write free-text notes for every call, repeating information already captured in the system | Simplify documentation — use structured fields and disposition codes instead of free-text paragraphs. If the call type and resolution are captured in drop-down fields, a free-text note should only be needed for exceptions |
| Too many disposition codes | Agents spend 30+ seconds scrolling through a list of 50+ codes to find the right one | Reduce to 15–20 primary codes organized by category. If agents cannot find the right code quickly, they either choose the wrong one or take too long choosing |
| Follow-up tasks | Agents send emails, create tickets, or update records after the call | Evaluate whether these tasks can be done during the call (while the customer confirms details) rather than after. If follow-up work is substantial, consider a dedicated back-office queue for post-call processing |
| Agent habit | Some agents use ACW as unofficial break time — staying in ACW status longer than needed | Track ACW duration by agent. If the team average is 45 seconds and one agent averages 120 seconds, that is a coaching conversation |
AHT by call type
A single AHT target for all call types creates distorted incentives. Simple calls (balance inquiry, password reset) should be much shorter than complex calls (technical troubleshooting, billing dispute). Setting a 5-minute target across all types pressures agents to rush complex calls and provides no useful signal on simple ones.
Example — AHT by call type for a customer support center:
| Call type | % of volume | Target AHT | Actual AHT | Gap |
|---|---|---|---|---|
| Account inquiry | 25% | 3 min | 2.5 min | On target |
| Billing question | 20% | 5 min | 5.5 min | Slightly over — investigate |
| Technical support (Tier 1) | 30% | 8 min | 9 min | Slightly over — may be appropriate if FCR is high |
| Complaint/escalation | 10% | 10 min | 12 min | Over — check if agents have authority to resolve |
| Order/change request | 15% | 4 min | 6 min | Over — likely a system or process issue |
The weighted average AHT for this operation is approximately 6 minutes. But the overall average is less useful than the per-type view, which shows exactly where the time problem is (order/change requests at 6 min vs. 4 min target) and where it is not (account inquiries are fine).
The AHT-FCR relationship
Every AHT reduction initiative should be evaluated against its impact on FCR. The two metrics must be tracked together.
| AHT trend | FCR trend | Interpretation | Action |
|---|---|---|---|
| Decreasing | Stable or increasing | Genuine efficiency improvement — agents are resolving faster without sacrificing quality | Continue the approach |
| Decreasing | Decreasing | Agents are rushing — shorter calls but more callbacks | Stop pushing AHT reduction. Investigate which call types are seeing FCR decline |
| Increasing | Increasing | Agents are taking more time and resolving more on the first call — total contact volume should be dropping | May be a net positive. Check whether total handle time (AHT × call volume including repeats) is decreasing |
| Increasing | Stable | Agents are taking longer without improving resolution — possible training gap or process complexity increase | Investigate by call type. Did a new product launch add complexity? Did a system change slow agents down? |
Total handle time is more meaningful than AHT alone:
Total handle time = AHT × (Total calls including repeat contacts)
An operation with a 5-minute AHT and 70% FCR handles 30% more total contacts than one with a 6-minute AHT and 85% FCR — because the repeat calls in the first operation add volume. The "efficient" operation may actually spend more total agent time per customer issue.
What not to do
Do not set AHT targets without FCR targets. An AHT target without a corresponding FCR floor tells agents that speed matters more than resolution. They will comply — and your callback rate will increase.
Do not use AHT as a primary performance metric for individual agents. AHT varies by call type, customer complexity, and factors outside the agent's control. An agent who handles mostly complaint calls will have a higher AHT than one who handles mostly account inquiries — and that is appropriate. Use AHT as a diagnostic metric at the team and call-type level, not as an individual performance score.
Do not compare AHT across different call centers without normalizing for call mix. A center that handles 60% technical support calls will have a higher AHT than one that handles 60% simple inquiries. The difference is call mix, not efficiency.
Do not reduce AHT by eliminating after-call documentation. ACW exists for a reason — accurate call records reduce repeat contacts, improve QA evaluation accuracy, and support dispute resolution. Eliminating documentation saves 30 seconds per call and costs far more in downstream problems.
Measuring the impact of AHT improvements
When you implement a change to reduce AHT, measure the full impact — not just whether AHT went down.
| Metric | What to track | Why |
|---|---|---|
| AHT by component | Did talk time, hold time, or ACW change? | Confirms the change affected the right component |
| FCR | Did first contact resolution change? | Catches quality degradation from rushing |
| Repeat contact rate | Are customers calling back within 24–72 hours at a different rate? | More sensitive than FCR for detecting unresolved issues |
| Customer satisfaction | Did satisfaction scores change? | Catches whether shorter calls feel worse to customers |
| Calls per agent per hour | Did throughput increase? | Confirms the AHT change translates to actual capacity gain |
| Occupancy | Did agent workload change? | If AHT dropped but occupancy stayed the same, the capacity gain was absorbed by volume increases or schedule changes |
A 30-second AHT reduction across an operation handling 10,000 calls per week saves 5,000 minutes of agent time per week — roughly 2.5 FTEs. That is meaningful. But only if FCR did not decline. If FCR dropped from 75% to 70%, the 5% increase in repeat contacts (500 additional calls per week at 6 minutes each) adds 3,000 minutes — erasing most of the gain.
