After-Call Work in Call Centers — What It Includes, Why It Takes Too Long, and How to Reduce It
After-call work (ACW) is the time an agent spends on tasks after the customer hangs up and before the agent becomes available for the next call. It is the third component of average handle time (AHT = talk time + hold time + ACW) and the one that is most often inflated by unnecessary process — yet least often analyzed.
In most call centers, ACW runs 30–90 seconds. But when it runs 2–3 minutes per call, the cumulative impact is significant. An agent handling 60 calls per day with 60 excess seconds of ACW per call loses 60 minutes of productive time — a full hour per day spent on post-call tasks that may not need to take that long. Across a 50-agent operation, that is 50 hours per day, or roughly 3 full-time equivalent agents worth of capacity consumed by ACW.
Reducing ACW is one of the highest-ROI improvements a call center can make because it directly increases agent availability without hiring, without changing call handling, and without affecting the customer experience. But reducing it requires understanding what agents are actually doing during ACW and why it takes as long as it does.
What ACW includes
ACW is not one task — it is a collection of post-call activities. The specific tasks vary by operation, but in most call centers they include some combination of:
| ACW task | What the agent does | Typical time |
|---|---|---|
| Disposition coding | Select a code from a list that categorizes the call type and outcome | 5–15 seconds |
| Call notes | Write a summary of the interaction — what the customer wanted, what was done, any follow-up needed | 15–60 seconds |
| System updates | Update the customer's account record — address change, plan modification, case status | 10–30 seconds |
| Follow-up creation | Create a ticket, task, or callback for an issue that could not be resolved on the call | 15–45 seconds |
| Email to customer | Send a confirmation, reference number, or summary of the resolution | 15–30 seconds |
| Wrap-up/personal time | Agent takes a brief pause before the next call — not an official task but often absorbed into ACW | 5–30 seconds |
Total typical range: 30–120 seconds for a well-designed process. Above 120 seconds consistently indicates a process problem, not a complexity problem.
Diagnosing why ACW is too high
Before reducing ACW, determine where the time is going. The cause determines the fix.
Step 1: Measure ACW by agent
Pull ACW duration by agent for the past 4 weeks. The distribution reveals whether the problem is systemic or individual.
| Pattern | What it means | Action |
|---|---|---|
| All agents averaging 90+ seconds | Systemic — the process or tools require too much time | Fix the process (see below) |
| Most agents at 45 seconds, 5 agents at 150+ seconds | Individual — specific agents are either struggling or using ACW as break time | Coaching for those agents |
| ACW varies significantly by call type | Some call types require more documentation than others | Set ACW targets by call type, not one blanket target |
| ACW increasing over time (was 45 seconds 6 months ago, now 75 seconds) | Something changed — new documentation requirement, new system, new process step | Identify what was added and whether it is necessary |
Step 2: Observe what agents do during ACW
Data shows how long ACW takes. Observation shows what agents are doing. Have a supervisor or QA analyst sit with 5–10 agents and watch what they do during ACW for 10–15 calls each.
Common findings:
| Observation | Root cause | Fix |
|---|---|---|
| Agent scrolls through 40+ disposition codes to find the right one | Too many codes, poorly organized | Reduce to 15–20 primary codes, organized by category |
| Agent types 3–4 sentences of free-text notes for every call | Documentation requirements are excessive or undefined | Define what must be documented vs. what is optional. Most calls need 1 sentence or none if disposition code captures the outcome |
| Agent re-enters information already visible on screen | System does not carry data between screens or fields | Fix system workflow — auto-populate fields that are already known |
| Agent switches between 3 applications to complete documentation | Multiple systems not integrated | Reduce the number of systems or create a unified post-call screen |
| Agent sits in ACW for 30+ seconds after completing all tasks | Using ACW as unofficial break time | Coach on ACW expectations, or provide dedicated recovery time between calls so agents do not need to use ACW for rest |
Reducing ACW: process changes
Simplify disposition codes
Disposition coding should take fewer than 10 seconds. If it takes longer, the code list is too long or poorly structured.
Before (50+ codes in a flat list):
Agents scroll through an alphabetical list of 50 codes, many of which are rarely used or overlap. Average time to find the right code: 15–25 seconds.
After (15–20 codes in categories):
| Category | Codes |
|---|---|
| Billing | Payment inquiry, Billing dispute, Plan change, Refund request |
| Technical | Troubleshooting, Outage report, Device issue, Installation |
| Account | Address update, Account inquiry, Cancellation, New service |
| General | Information request, Complaint, Escalation, Follow-up needed |
Agents select a category (3 seconds), then select the specific code (3 seconds). Total: 6 seconds.
If agents frequently select "Other" or "General inquiry," your code list does not match your actual call types. Audit "Other" entries monthly and add specific codes for the most common entries.
Define documentation standards
The most common ACW inflator is undefined documentation requirements. When agents are told to "document the call" without clear guidance on what that means, some write nothing and some write paragraphs.
Define what must be documented:
| Scenario | Required documentation | Example |
|---|---|---|
| Call resolved, routine issue | Disposition code only — no notes needed | Code: "Payment inquiry" |
| Call resolved, non-routine issue | Disposition code + 1 sentence describing what was done | "Credited $45 for billing error on March statement" |
| Call not resolved, follow-up needed | Disposition code + ticket/task + details for the follow-up agent | "Customer reports intermittent outage since March 10. Escalated to Tier 2, ticket #4892" |
| Call involved a customer complaint or dispute | Disposition code + specific notes on what was promised | "Waived late fee per customer request. Confirmed no future fees will be waived" |
This structure tells agents exactly what level of documentation each call type requires. Routine calls that were resolved need only a code click — no typing. Only exception cases require written notes.
Move work into the call
Some ACW tasks can be completed while the customer is still on the line — during natural pauses in the conversation or while the customer confirms details.
| Task | During-call approach |
|---|---|
| Disposition coding | Select the code as the resolution becomes clear, before the call ends |
| System updates | Make the account change while talking to the customer ("I'm updating that now...") |
| Confirmation email | Trigger the email before ending the call ("You'll receive a confirmation in a few minutes") |
| Call notes | Type the key detail during the conversation, not after |
Moving 30 seconds of post-call work into the call adds 30 seconds to talk time but eliminates 30 seconds of ACW. The net effect on AHT is zero — but the agent becomes available for the next call immediately, which improves throughput during high-volume periods because the agent spends zero time in a non-available state.
Address ACW as break time
In many call centers, agents use ACW as unofficial recovery time between calls — especially when occupancy is high and agents are handling calls back-to-back. The agent finishes the post-call tasks in 30 seconds but stays in ACW status for an additional 30–60 seconds to catch their breath.
This is a symptom, not a cause. If agents need ACW for rest, the real issue is that occupancy is too high and there is no built-in recovery time. Cracking down on ACW duration without addressing occupancy will push agents to find other ways to get breaks — longer hold times, extended talk time, or increased absenteeism.
Solutions:
- If occupancy is chronically above 85%, add agents to bring it to the 75–85% range. Agents will have natural idle time between calls and will not need to manufacture it in ACW
- Some operations build in a short auto-ready delay (10–15 seconds after ACW before the next call routes). This provides a structured micro-break without agents extending ACW manually
Measuring the impact of ACW reduction
When you implement changes to reduce ACW, track the full impact — not just whether ACW went down.
| Metric | What to track | Why |
|---|---|---|
| ACW duration (average and by agent) | Did it decrease? By how much? | Confirms the change is working |
| AHT | Did total AHT change? | If ACW went down but talk time went up by the same amount (agents moved work into the call), net AHT is unchanged — which is fine, throughput still improved |
| Calls per agent per hour | Did throughput increase? | The real test — agents should handle more calls if ACW decreased |
| Documentation quality | Are call notes accurate and complete? | Catches whether ACW reduction degraded documentation — audit a sample of records weekly |
| QA scores | Did quality change? | Ensures that faster post-call work is not cutting corners on process compliance |
| Repeat contact rate | Are customers calling back at a higher rate? | If follow-up tasks are being skipped or done poorly due to rushed ACW, repeat contacts will increase |
The capacity math
ACW reduction translates directly to increased agent availability. Here is what different reductions mean for a 100-agent operation handling 60 calls per agent per day:
| ACW reduction per call | Daily time saved per agent | Daily time saved (100 agents) | Equivalent FTEs |
|---|---|---|---|
| 15 seconds | 15 minutes | 25 hours | 1.5 FTEs |
| 30 seconds | 30 minutes | 50 hours | 3 FTEs |
| 45 seconds | 45 minutes | 75 hours | 4.5 FTEs |
| 60 seconds | 60 minutes | 100 hours | 6 FTEs |
A 30-second ACW reduction across a 100-agent operation is equivalent to hiring 3 additional agents — at zero incremental cost. At a fully loaded cost of $30,000–$40,000 per agent per year, that represents $90,000–$120,000 in avoided hiring cost annually.
This is why ACW is one of the first places to look when an operation needs more capacity but cannot hire. The capacity may already exist — trapped in unnecessary post-call process.
What not to do
Do not eliminate ACW entirely. Some operations set ACW to zero — agents are auto-routed to the next call the moment the previous call ends. This eliminates documentation, guarantees errors in call records, and creates an environment where agents feel they have no control over their workflow. Every call needs at least 10–15 seconds for disposition coding.
Do not set an ACW timer that forces agents out. A 30-second forced timer sounds efficient but creates problems when a call genuinely requires 60 seconds of follow-up work. The agent either does the work poorly (to beat the timer) or switches to an unavailable state to finish it (which defeats the purpose). Set ACW expectations and coach to them rather than enforcing hard cutoffs.
Do not reduce ACW without checking documentation quality afterward. Fast ACW with no documentation is worse than slow ACW with good documentation. The downstream cost of missing or inaccurate call records — failed follow-ups, repeat contacts, disputes with no record — exceeds the time saved by eliminating documentation.
