Average Handle Time is one of the most measured metrics in the contact centre industry. It is also one of the most misunderstood, most misapplied, and — depending on how it’s used — one of the most damaging.
I’ve been working in contact centres for nearly two decades. I can count on one hand the organizations I’ve seen use AHT correctly. I’ve lost count of the ones I’ve seen use it to inadvertently destroy their customer experience and demoralize their teams simultaneously.
This post is my attempt to set the record straight on what AHT actually is, what it’s actually good for, and — critically — what it should never be used for.
What AHT Actually Measures
Average Handle Time measures the average total duration of a customer interaction, from the moment the agent picks up to the completion of all post-call work. It has three components:
Talk time — the live conversation between agent and customer. Hold time — time the customer spends waiting while the agent looks something up, consults a colleague, or processes a request. After-call work (ACW) — the wrap-up tasks after the customer hangs up: case notes, CRM updates, follow-up actions.
The formula is straightforward:
AHT = (Total Talk Time + Total Hold Time + Total After-Call Work) ÷ Total Calls Handled
It is the bedrock for all contact centre planning systems, as well as a key component of all Erlang calculations — unless you know this metric you cannot plan effectively. That’s the correct framing. AHT is a planning input. It tells you how much time an agent spends per interaction, which determines how many agents you need to handle a given volume at a given service level.
That’s what it’s for. That’s all it’s for.
The Problem: What AHT Gets Used For Instead
Many contact centres use AHT as a personal agent target — in fact, in a number of surveys, it is consistently cited as the most common metric in contact centres — although this is normally deemed a “bad thing.” The problem is that it tends to make agents focus on speed of service rather than quality of service. It also tells you nothing about the outcome of a call. It tends to drive a behaviour of agents trying to manipulate the AHT.
I’ve seen every variant of this manipulation in practice:
Agents wrapping up calls before the customer’s issue is fully resolved — technically within target, practically useless. Agents not capturing complete notes because wrap time is measured against a target. Agents rushing customers through scripted responses without pausing to check whether the answer actually solved the problem. In every case, the AHT metric looked fine. The customer experience was quietly eroding.
At a contact centre for a major US brand, a management advisory firm advised the company to maintain an AHT of 480 seconds based on a benchmark study. Given the nature of the business and the typical problems customers had, calls would frequently extend beyond this threshold. Because of the company’s strict adherence policy, agents routinely received negative performance reviews — for doing their jobs correctly.
This is the failure mode. The target becomes the goal rather than the outcome the target was supposed to represent.
The Commitment Gap
The most expensive version of AHT misuse is when leadership uses it to set service commitments without checking whether the operation can actually fulfill them.
This is surprisingly common. An executive commits to a 30-minute email response time — or an 80/20 resolution model where 80% of calls are resolved in 20 seconds or less — because it sounds right in a contract negotiation or a board presentation. What nobody asks is: at what staffing level, at what hours, against what expected volume, using what tools, is this commitment actually achievable?
The commitment is made in a boardroom. The consequences arrive in a queue.
If you’re going to commit to a 30-minute email response, you need to know precisely what happens to that target when two agents are out sick, volume spikes after a product release, and it’s a Monday morning in peak season. Building the staffing model around realistic forecasts rather than aspirational commitments is the only way to make a service promise you can keep.
The tool that actually answers the staffing question — for voice channels — is Erlang C modelling. It gives you the mathematical relationship between incoming volume, handle time, and the number of agents required to meet a service level target. AHT is an essential input into that calculation. But it’s an input — not the output. The output is a staffing requirement. Most organisations have this exactly backwards.
Who Is Your Customer — And What Do They Actually Need?
One of the things I’ve always pushed back on in AHT conversations is the implicit assumption that all calls are equivalent. They’re not.
Most contact centres use templates and structured call flows to ensure consistency. This works for common, low-complexity issues. It falls apart the moment the customer in front of you doesn’t fit the template — because their problem is more complicated, because their account is more complex, because they’re frustrated after a previous interaction that didn’t resolve their issue.
The agent’s job in the first two minutes of any call is to understand what kind of call this actually is. Is this a quick account enquiry that should be in and out in three minutes? Or is this a complex billing dispute with a premium customer that needs time, empathy, and the authority to go off-script to reach a real resolution?
If an agent is watching their handle time target while trying to make that judgment, you’ve created a direct conflict between serving the customer and protecting their performance metric. This pits the agent’s interests squarely against the customer’s interests — and demoralizes smart agents who would otherwise be capable of resolving issues efficiently.
The goal should always be to give the customer the level of service their situation requires. Sometimes that’s fast. Sometimes it isn’t. A rigid AHT target that penalizes agents for the latter is a structural failure in your performance management design.
AHT in an AI-Augmented Environment
The AHT conversation is changing significantly as AI becomes part of the operational picture.
When AI surfaces relevant knowledge base articles in real time during interactions, agents spend less time searching for information and more time actually helping the customer. The result is often a genuine reduction in AHT — but achieved through better capability rather than through rushing. That’s the right kind of AHT improvement.
According to IBM’s 2026 contact centre automation analysis, one bank that introduced an AI-driven assistant achieved a 6% reduction in average handle times along with lower training requirements — without scripting a single new call flow. That’s worth noting: the AHT reduction came from capability improvement, not from target pressure.
AI is also changing what the remaining human interactions look like. As routine, low-complexity queries are increasingly handled by automation and self-service, the contacts that reach human agents are skewing toward the more complex, more time-consuming, and more emotionally charged. In this environment, a rising AHT may not indicate inefficiency — it may indicate that your AI deflection is working and your agents are handling the work that genuinely requires them.
A higher AHT isn’t automatically a problem. Sometimes it simply means your team is handling harder work — more consultative conversations, more complex escalations, more compliance-sensitive calls. If you’re measuring AHT without understanding what’s driving it, you’re looking at a number without the context that makes it meaningful.
What to Measure Instead (or Alongside)
AHT tells you how long interactions take. It doesn’t tell you whether they were successful.
The metrics that tell you whether interactions were successful are:
First Contact Resolution (FCR) — was the customer’s issue resolved without them needing to contact you again? Companies that have removed AHT as an agent performance metric have often found that, after an initial rise in AHT, FCR rates tend to improve. The two metrics pull in opposite directions when you use both as agent targets. FCR usually wins on what actually matters.
Customer Satisfaction Score (CSAT) — how did the customer rate their experience? The CX Leader’s KPI Playbook has the full framework for how CSAT should be used — as a coaching tool and a signal, not as the sole measure of agent performance.
Repeat Contact Rate — how many customers contacted you more than once about the same issue? This is the operational consequence of low FCR and the metric that most clearly connects AHT gaming to actual customer outcomes. A short call that doesn’t resolve the issue costs you two calls.
Net Promoter Score (NPS) and Customer Effort Score (CES) — both are more meaningful than AHT for understanding whether your operation is delivering the experience your customers actually want.
The real benchmark for improved AHT is reduced resolution times plus high levels of customer satisfaction. Improving AHT means reducing the time it takes to get to a resolution while maintaining or improving quality — not simply cutting calls short.
Should You Get Rid of AHT?
No. And this is important to say clearly, because the argument I’ve been making might sound like a case for abandoning the metric entirely. It isn’t.
AHT is genuinely useful. Its singular primary purpose must be to forecast the demand for agents — AHT should mainly focus on planning capacity properly and scheduling agents with precision. It should also be used to identify flaws in the contact centre’s processes, knowledge, infrastructure, and systems that hinder agents from providing the highest level of service and lead to unnecessarily lengthy calls.
Those are legitimate and important uses. A team where agents spend all day on a single call is not serving its customers or its business. Understanding average handle time by queue, by issue type, by channel, and by agent helps you identify where training is needed, where process is broken, and where tools are inadequate.
What AHT should not be is an individual performance target that agents are evaluated against and stressed about during interactions. The moment you put a timer on an agent’s dashboard and tell them their performance review depends on it, you’ve created an incentive structure that works against the customer.
Use AHT to plan. Use it to identify process problems. Use it to understand your operation. Don’t use it to measure individual agents, and don’t let it drive the service standards you communicate to customers.
The Practical Takeaway
If AHT is currently used as an individual agent target in your operation, I’d challenge you to run a simple experiment: remove it as a personal metric for one team for 90 days, while continuing to track FCR and CSAT. In my experience — and this is backed by the industry research — you’ll see AHT rise slightly in the first few weeks as agents stop rushing, and then watch your repeat contact rate fall and your satisfaction scores improve.
The math that seemed to support AHT as a performance metric — more calls handled per agent per day — turns out to be wrong when you factor in the cost of repeat contacts, the attrition driven by demoralized agents, and the churn driven by customers who didn’t get their problem solved the first time.
You’re not saving money by rushing calls. You’re spending it somewhere else — and somewhere more expensive.
Hutch Morzaria is a Director-level CX and Support Leadership professional with 19 years of experience building global support organizations across SaaS, Fintech, and enterprise technology. He has hired dozens of support and CX leaders across his career and holds ITIL Expert certification across V3 and V4.
