Skip to content

Bananaflow MCP is live.

Read the docs

AHT Full Form: Formula, ACW and Benchmarks

Jaseem Abbas, founder of Bananaflow, in glasses and a dark top against a pale background.Jaseem Abbas

Published
Updated
Cover for "AHT Full Form: Formula, ACW and Benchmarks": a white isometric line diagram of one average across many calls: five towers in a row, each as tall as its call and none the same, beside a pill reading Each call, with one bold outline lying across all five at the height they average out to, its corner marked by a lit pill reading AHT, drawn over a soft, out-of-focus periwinkle in muted aqua, its flat pinwheel petals blurred into veils at both ends, with a pale blush petal glow at the left end and a soft peach glow at the right.

AHT full form is Average Handle Time (also written average handling time): the average time an agent spends on one call, from pick-up to the moment they are ready for the next one. In a BPO, it counts three things: talk time, hold time and after-call work. In Hindi you could say "ek call par lagne wala ausat samay".

Here is why it matters. A team lead at an insurance renewal helpdesk in Pune opens the daily report and sees AHT at 6.0 minutes. Her manager wants 5.5. She has 40 agents. She can cut after-call work and get there, or she can tell agents to finish calls faster and watch the same customers ring back the next morning. The number is the same either way. The cost is not.

This guide covers the AHT meaning in a call center, what it includes, what ACW is, the formula with a worked example you can check, what a good AHT looks like, and how to lower it without hurting resolution. The Pune numbers are an example we made up for this post, not customer data.

What is the full form of AHT?

AHT stands for Average Handle Time, and in a call center it is the average time one call takes from start to finish, including wrap-up. Some floors say average handling time. It is the same measure.

Two notes keep this page focused. First, AHT has other meanings outside call centres, in medicine and in chat tools, and this guide is only about voice calls. Second, chat needs its own definition, because an agent can handle several chats at once. NICE, a contact center software company, says exactly this in its glossary (NICE on average handle time).

If you came from the wider picture, our guide to call center metrics puts AHT next to first call resolution, CSAT, abandonment and occupancy. This page goes deeper on AHT alone. To see AHT on a screen with the other numbers, read what to put on a call center dashboard.

What does AHT include?

AHT includes talk time, hold time and after-call work. It leaves out the time before the agent picks up and the idle time after the agent is ready again.

  • Part: Talk time

    In AHT?
    Yes
    What it is
    Minutes the agent and the caller are speaking
  • Part: Hold time

    In AHT?
    Yes
    What it is
    Minutes the caller waits while the agent looks something up or consults someone
  • Part: After-call work (ACW)

    In AHT?
    Yes
    What it is
    Wrap-up after the caller hangs up: notes, outcome code, system updates
  • Part: Ring time

    In AHT?
    No
    What it is
    Time the phone rings before the agent answers
  • Part: Queue wait

    In AHT?
    No
    What it is
    Time the caller waits for an agent. It shows up in average speed to answer and abandonment
  • Part: Idle time

    In AHT?
    No
    What it is
    Time the agent is logged in and ready, waiting for the next call

One more thing worth saying. AHT is not the same as call duration. Call duration usually stops when the line closes. AHT keeps counting through the wrap-up, so a call that lasts 4 minutes on the phone can have a handle time of 5.

How AHT is built. Talk, hold and after-call work are added together across every handled call, then divided by the number of calls.

What is ACW in a call center?

ACW means after-call work: the tasks an agent finishes after the caller has hung up and before they are ready for the next call. If you searched "ACW full form" and landed on power-plant pages, those are about a different ACW (cooling water). In a BPO, ACW is wrap-up time. AI call summaries can take over part of that work; see AI call summaries.

What fills ACW depends on the process. Common tasks are:

  • Choosing a disposition code, which is the outcome code for the call, such as "renewal confirmed" or "callback requested".
  • Writing a note in the CRM about what the caller said.
  • Updating a ticket or a record.
  • Raising a follow-up task, such as a request to send a document.

ACW has a reputation for hiding other things. Agents sometimes use wrap-up as a short breather between hard calls. That is human, and a hard cap on ACW tends to backfire in the same way a hard cap on AHT does. A better approach is to ask why wrap-up takes long: a clumsy form, too many fields, notes typed from memory? Fix that and the time falls without a fight.

How do you calculate AHT?

You calculate AHT by adding total talk time, total hold time and total after-call work over a period, then dividing by the number of calls handled. NICE gives the same voice formula: "AHT = (total talk time + total hold time + total after-call work) ÷ number of handled calls."

Here is a worked example. An insurance renewal helpdesk in Pune handled 2,400 calls in one day. This is an example we made up so the arithmetic can be checked.

  • Part: Talk

    Per call (minutes)
    4.5
    Calls
    2,400
    Total minutes
    10,800
  • Part: Hold

    Per call (minutes)
    0.6
    Calls
    2,400
    Total minutes
    1,440
  • Part: After-call work

    Per call (minutes)
    0.9
    Calls
    2,400
    Total minutes
    2,160
  • Part: Total handle time

    Per call (minutes)
    6.0
    Calls
    2,400
    Total minutes
    14,400

The sums:

  1. Talk: 2,400 × 4.5 = 10,800 minutes.
  2. Hold: 2,400 × 0.6 = 1,440 minutes.
  3. After-call work: 2,400 × 0.9 = 2,160 minutes.
  4. Total handle time: 10,800 + 1,440 + 2,160 = 14,400 minutes.
  5. AHT: 14,400 ÷ 2,400 = 6.0 minutes, which is 360 seconds.

Keep one rule for teams. Team AHT is total handle time divided by total calls. Do not average each agent's own AHT, because an agent who took 20 calls would count the same as one who took 200. Knowmax, an Indian knowledge-management vendor, gives the same advice in its FAQ: for a team figure, divide total time by total calls and do not average each agent's AHT (Knowmax on average handle time).

How do you calculate AHT in Excel?

Put one row per call in a sheet: a call ID in column A, then talk seconds in B, hold seconds in C and after-call work seconds in D. Then one formula gives AHT in seconds:

=SUM(B2:D2401)/COUNTA(A2:A2401)

Divide by 60 for minutes. The range here is 2,400 calls, rows 2 to 2401. Change the last row to match your own export. If your dialer report gives totals instead of rows, add the three totals and divide by the calls handled, as in the Pune table.

What does AHT mean for staffing and cost?

AHT turns directly into seats. Total handle time is the work; occupancy says how much of a shift an agent can spend on it. Look at the Pune day again.

  1. 14,400 minutes ÷ 60 = 240 agent-hours of work.
  2. An 8-hour shift at 75% occupancy gives 6 productive hours per agent (8 × 0.75 = 6).
  3. 240 ÷ 6 = 40 agents.

The 75% matches the occupancy figure in our call center metrics guide. Real floors also staff for peaks and absences, so treat 40 as the base load, not the full roster. Our Erlang C guide works the same Pune day half hour by half hour, with shrinkage and shifts.

Now cut after-call work from 0.9 to 0.4 minutes per call, for instance by filling notes automatically:

  1. After-call work: 2,400 × 0.4 = 960 minutes.
  2. Total handle time: 10,800 + 1,440 + 960 = 13,200 minutes.
  3. AHT: 13,200 ÷ 2,400 = 5.5 minutes.
  4. Hours: 13,200 ÷ 60 = 220, and 220 ÷ 6 = 36.7, so about 37 agents.

That is three seats from half a minute of wrap-up, and 1,200 minutes saved (14,400 minus 13,200). For what a seat and a minute cost in rupees, see our guide to call center cost per call in India.

What is a good AHT?

A good AHT depends on the type of call, and no single number fits every floor. We found no public Indian benchmark that we could verify, so the table below uses the sources we could check, each labelled by quality.

  • Figure: Industry benchmark of 7 minutes, varying by call type and industry

    Source
    SQM Group, 7 essential customer service metrics (December 2022)
    How far to trust it
    Quoted from the page. SQM benchmarks North American centres, so it is a rough guide for India
  • Figure: Around 4 minutes of AHT (60-minute rolling mean, 233 to 252 seconds) and about 75 seconds of ACW (73 to 78 seconds) across 1.2 million calls

    Source
    Emory Healthcare call center, US, JMIR Medical Informatics 2026
    How far to trust it
    A published study, so the data is real, but it is one healthcare centre in the US. The excerpt does not say whether AHT there includes ACW
  • Figure: Retail 3 to 4 minutes, banking 4 to 6, telecom 5 to 7, tech support 8 to 10, insurance 5 to 7

    Source
    One vendor's blog, which calls them estimates and gives no source
    How far to trust it
    Weak. Shown only because readers look for it. We would not set a target from it

What you can take from this is the shape, not a number. Numbers across industries range from about 4 minutes to 10, and the same call type can differ by a minute between centres. So set your target this way:

  1. Split calls by type: renewal, balance enquiry, complaint, new sale.
  2. Take your own last 90 days for each type.
  3. Read the median and the 80th percentile, not just the average.
  4. Aim to improve each type a little, while watching first call resolution.

The Emory study, published in JMIR Medical Informatics, and the SQM Group metrics page are the two sources worth opening if you want to read further.

Why is AHT important in a BPO?

AHT matters in a BPO because it drives staffing, and in many contracts it drives price. If a client pays per minute or per seat, a floor whose AHT drifts from 6.0 to 6.5 minutes needs about 8% more agent time for the same calls (0.5 ÷ 6.0). A BPO that cannot say what its AHT is by call type cannot quote a fair price or plan a shift.

The trap in a BPO is turning AHT into a score for each agent. Ranking people on it teaches them to end calls, not solve them. Our guide to how to handle angry customers on calls gives the opposite case: some calls should run long, and the agent who stays on the line until the caller is calm saves a complaint.

Why does AHT alone mislead?

AHT alone misleads for two reasons: it is a skewed average, and it can fall while the real cost rises.

Is the average the right number?

Not always. Call lengths are skewed: many short calls and a few very long ones. A few long calls pull the average up. An illustration, with made-up numbers: nine calls of 3 minutes and one call of 33 minutes give a mean of 6.0 minutes ((9 × 3 + 33) ÷ 10) while the median call is 3 minutes. The CEO of CARE4CC, quoted by Call Centre Helper, said the average call duration "is heavily distorted by the outliers" (Call Centre Helper on how AHT is distributed). Track the median and the share of calls over a limit, such as 15 minutes, alongside the mean.

Can lower AHT cost you money?

Yes. The saving shows up in the AHT report. The cost shows up in next week's call volume. SQM Group says that when AHT moves, "it is critical to evaluate its impact on FCR and Csat" (SQM Group). NICE puts it this way: pressure to end calls quickly "can increase repeat contacts, transfers and customer frustration", while a longer call may be efficient if it resolves a complex issue completely.

Back to Pune. Say the team leader pushes agents to cut after-call work, and as a side effect first call resolution (FCR) falls by 5 points. The 5 points is our assumption for the example.

  1. Repeat calls: 5% of 2,400 = 120 more calls a day.
  2. Extra minutes at 6.0 minutes each: 120 × 6 = 720 minutes.
  3. Saved minutes from the shorter wrap-up: 1,200.
  4. Net saving: 1,200 − 720 = 480 minutes, which is 8 agent-hours, or about 1.3 agents (8 ÷ 6).

So most of the saving is gone, and the customers who had to ring twice are less happy. If FCR fell by 10 points, the repeats would cost 1,440 minutes and the change would lose money.

The AHT trade-off in the Pune example. Cutting after-call work saves 1,200 minutes, but a 5-point drop in first call resolution adds 120 repeat calls and 720 minutes back.

How big is FCR in practice? SQM Group reports that the average call center FCR is 71%, with 70 to 79% counted as good and 80% and above as world class, and says that each 1% gain in FCR cuts operating costs by about 1% (SQM Group on first call resolution). That last point is SQM's own research claim, so read it as a guide, not a law. For how to define and measure FCR, see our FCR section.

How do you reduce AHT without hurting resolution?

You reduce AHT by removing the causes of long handling, one component at a time, and by checking FCR every time. The list below follows the three parts of the formula.

Talk time

  • Route callers to the right person the first time, so the agent does not spend three minutes before saying "this is not my team".
  • Put customer details and the answer on one screen, so the agent is not searching.
  • Cut duplicate verification: asking for the same details twice is the most common waste.
  • Coach with real calls. Our call center training manual covers how to build that into a week.

Hold time

  • Ask where agents put callers on hold and why. Usually it is one slow lookup, or a question only a supervisor can answer. Fix that lookup or give the agent the authority.
  • Use a warm call transfer with a short summary when a call has to move, so the second agent does not start from zero. Choosing between a warm and a cold transfer changes how long the caller waits. Our call center scripts include hold and transfer lines you can copy.

After-call work

  • Remove fields that no one reads.
  • Fill dispositions and notes automatically where you can, from the call itself.
  • Give agents a fixed, short list of outcome codes instead of a long drop-down.

NICE's list of responsible ways to reduce AHT is similar: better routing, context and knowledge on one screen, fewer duplicate verification steps, automated summaries and dispositions for wrap-up, fixing processes that cause holds and transfers, and moving routine tasks to self-service.

What not to do:

  • Do not set an AHT target for each agent with a bonus attached.
  • Do not let agents transfer hard calls just to close them.
  • Do not skip verification to save 20 seconds.
  • Do not cap ACW with a timer without finding out why it is long.

Change one thing at a time and check the median and FCR a week later, as we suggest for all call center metrics.

How do AI voice agents change AHT?

An AI voice agent changes AHT in two ways, and neither is a promise.

First, the mix changes. If an agent takes the routine calls, the human floor is left with the harder ones. Its AHT per call can go up while total cost goes down, so compare cost per resolved issue, not AHT alone. Our guide to telecaller versus AI calling agent cost works through the cost side.

Second, the wrap-up work can be done by the system. After each call, Bananaflow keeps a transcript, a recording, the minutes used and the extracted variables. A QA step can return tags, sentiment and a summary, and the call outcome is recorded as a disposition, which is an outcome code. These are the items a person would type in ACW. The results can go to your own system by webhook after each call, and the reports show KPIs and trends with CSV export (call QA, analytics).

How Bananaflow fits

When a call has to go to a person, transfer works on Twilio, Plivo, Telnyx, Cloudonix and Asterisk, and on Twilio the person hears a short summary of the call so far. Check that your carrier is on that list.

What should you remember about AHT?

AHT full form is Average Handle Time: talk, hold and after-call work, divided by calls handled. Work it out for your own calls, split by call type, and read the median beside the average. Turn it into seats to see what it costs, as in the Pune example (14,400 minutes is about 40 agents). Compare any figure from outside with its source before you use it as a target. And never cut AHT without watching first call resolution, because a repeat call costs a whole call again.

FAQ

Frequently asked questions

  • What is AHT in BPO?

    AHT in a BPO is Average Handle Time: the average time an agent spends on one call, counting talk, hold and after-call work. BPOs use it to plan seats, price work by the minute and set targets for each process. It should never be read alone, because a low AHT with many repeat calls is a worse result than a higher AHT that resolves the issue.

  • What is a good AHT in a call center?

    It depends on the call type, so there is no single good number. SQM Group gives 7 minutes as an industry benchmark for North American centres, and one vendor guess puts retail at 3 to 4 minutes and tech support at 8 to 10. We found no verified public benchmark for India. Set your target by call type from your own last 90 days, and check that first call resolution holds.

  • Is hold time part of AHT?

    Yes. The common formula adds talk time, hold time and after-call work, then divides by the calls handled. If you want a place to start cutting AHT, start with hold time: find the lookups that make agents put callers on hold and fix those first.

  • Does AHT include ring time or queue time?

    No. Time spent ringing and time the caller waits in the queue are left out. They show up in other measures, such as average speed to answer and abandonment. AHT covers only the time an agent spends on the call and the wrap-up after it.

  • What is the difference between AHT and ATT?

    ATT is average talk time: only the minutes the agent and caller speak. AHT is larger, because it adds hold time and after-call work. In the Pune example in this guide, talk time is 4.5 minutes per call and AHT is 6.0.

  • How can an agent reduce AHT?

    Learn where the system holds the answer so you do not put callers on hold, confirm the issue early, and write notes while you talk so after-call work is short. Do not rush the caller or skip verification. If the caller rings back, the saved minutes are gone, and your first call resolution suffers.

  • Can an AI voice agent lower AHT?

    Here is how it works: after each call you get a transcript, a recording, a summary and a call outcome code, which is the work a person would otherwise type in the wrap-up. Whether that cuts your AHT is something to measure on your own calls.

Written by

  • Jaseem Abbas, founder of Bananaflow, in glasses and a dark top against a pale background.

    Jaseem Abbas

    Founder, Bananaflow

    Jaseem founded Bananaflow to build voice AI that understands how people in India really speak, including the switch between languages in the middle of a call.

Similar articles

Bring us one process. We'll pilot it on your real calls, with your team as the fallback.

Plan a pilot