Average handle time is how long an agent spends on one ticket from start to finish. That includes the conversation, any time the customer waits, and the notes and tagging afterwards. The quickest way to bring it down is to take work off the agent’s plate — saved replies, easy-to-find answers, better routing — not to hurry the customer.
On this page
- What is average handle time, and what is a good number?
- Why pushing agents to be faster backfires
- How to reduce average handle time
- Which one should you try first?
- Four things not to do
- How do you know if it worked?
- How LiveAgent helps
- Where to start this week
What is average handle time, and what is a good number?
Average handle time is the average time an agent spends on one ticket. It is not just the chat or the call. It adds up three things:
| Part of the ticket | What it covers | Where the time goes missing |
|---|---|---|
| Talking time | The actual chat or call with the customer | Asking for details the customer already gave you |
| Waiting time (hold time) | Any moment the customer sits there waiting | Checking a policy, asking a colleague, waiting for a slow screen |
| Wrap-up time (after-call work) | Notes, tags, updating other systems, follow-up emails | Writing the same summary by hand, every time |
Look at that list again. Two of the three have nothing to do with how fast anyone talks.
That is the most useful thing to know about this number, and it shapes everything below.
A good target depends on your industry and on how people get in touch. The same six-minute call could be perfectly normal at a bank and slow at an online shop, because one has identity checks it cannot skip and the other does not. So treat any single number that claims to cover every industry as noise, and compare yourself against benchmarks for your own industry instead.
Two things to know before you go looking for a target. Email usually has no handle time figure at all, because email gets answered whenever someone reaches it, so teams measure response time instead. And the trustworthy per-channel numbers tend to sit inside paid research, such as ContactBabel’s 2026-27 US Contact Center HR & Operational Benchmarking Report , which surveys 207 US organizations. Free roundups mostly quote each other.
Why pushing agents to be faster backfires
Handle time tells you what a ticket cost you. It never tells you whether the customer got their problem solved.
Here is the pattern to watch for. If handle time drops and the number of issues solved on the first try drops with it, your team is closing tickets before the problem is fixed. That work does not go away. It comes back as a second ticket, from someone who is now annoyed, and that second ticket takes longer than the first one would have.
So never chase this number on its own. Keep it next to two others: how often you solve the issue on the first try (first contact resolution), and how happy customers say they are. If either of those slips, stop.
Ready to grow your business?
Start your free trial today and see results within days.
How to reduce average handle time
There is no single trick to how to reduce average handle time. But everything below has one thing in common: it takes work away from the ticket instead of taking time away from the customer.
1. Start with the wrap-up, not the conversation
The minutes after a ticket closes are the easiest to win back, and almost nobody measures them. Agents write the same summary again and again, tag things three different ways, and copy the same details into a second system by hand.
This is not just our opinion. Natterbox’s State of the Contact Center 2026 , which looked at 58.2 million calls from live phone records rather than survey answers, puts it plainly: “Talk time is not the constraint. Wrap-up, logging, QA, and follow-up are.”
So go and watch. Sit with three agents and write down every step between finishing one ticket and starting the next. Teams often find a minute or more of pure admin per ticket that nobody knew was there. Ready-made summaries, a short tag list people can actually remember, and letting your systems talk to each other will clear most of it. You will not have touched a single conversation.
2. Write saved replies for your twenty most common questions
Pull the last three months of tickets and group them by what people asked. Take the top twenty. Right now your team is typing those answers from scratch, slightly differently, several times a day.
Saved replies fix that. One rule though: leave clear gaps for the agent to fill in, so the reply still sounds like a person. A saved reply that reads like a form letter just creates another question.
3. Get the answers out of people’s heads
Waiting time is where “only Sarah knows how to do that” shows up as a cost. If two people understand refunds on old plans, every refund ticket either waits for them or gets passed along.
An internal knowledge base solves it, as long as agents can search it without leaving the screen they are working in. A wiki in another browser tab does not help, because stopping to go and look is the thing that was slowing them down.
4. Send the ticket to the right person the first time
Every ticket that lands in the wrong queue costs you twice: the time of the person who cannot help, plus the time of the person who can. The customer explains everything twice, too.
Good routing rules and a short phone menu cost far less than any training program. Keep the menu short. Long menus look tidy on paper and make people hang up, which moves your problem somewhere you are not looking.
5. Have the customer’s details ready before anyone says hello
If your agent has to ask for an order number the customer already typed into a form, the waiting has started before the conversation has. Put the account details, recent tickets, and anything the customer filled in on the same screen as the reply box.
Ticket validation helps as well. It catches empty, duplicate, or clearly invalid requests before an agent ever opens them.
6. Fix the three things that cause your slowest tickets
Some tickets are slow because something upstream is broken. A confusing invoice, a cancellation process nobody understands, an integration that keeps failing. No saved reply will shorten those.
Sort your ticket types by the total time they eat, not by how many there are. Take the worst three to whoever owns that part of the business. This is the only item on the list that removes tickets instead of shortening them, and it is the one support teams skip.
7. Look at your slowest tickets, not your averages
An agent’s average handle time mostly tells you what kind of tickets they get. Someone handling billing disputes will always look slower than someone resetting passwords. Coaching on that average is unfair and it does not work.
Read your slowest tickets instead — say the slowest one in ten. They will show you the saved reply that is missing, the help article that is wrong, and the process that forces an agent to escalate. Every fix you find in there helps the whole team.
Where AI helps, and where it does not
AI is genuinely good at the two parts that cost the most: the wrap-up and looking things up. An agent editing a draft reply is faster than one staring at an empty box, and an automatic summary of the conversation removes the note-writing step.
Two warnings. AI that answers customers on its own does not lower handle time — it removes the quick, easy tickets from your queue, which actually pushes your average up even though you are doing less work overall. And AI drafting from a thin knowledge base writes confident nonsense that your agents then have to unpick, which costs more time than typing the answer themselves. Sort the knowledge base out first.
Which one should you try first?
Pick the row that sounds like your team.
| If this sounds familiar | Start here | How hard | Time you get back |
|---|---|---|---|
| Agents pause between tickets | Fix the wrap-up (1) | Easy | Wrap-up time |
| The same questions come in daily | Saved replies (2) | Easy | Talking time |
| Agents keep asking colleagues | Internal knowledge base (3) | Medium | Waiting time |
| Tickets get passed around | Routing and phone menu (4) | Medium | Talking and waiting |
| Customers repeat themselves | Details ready up front (5) | Medium | Talking time |
| A few ticket types eat your week | Fix the cause (6) | Hard | Whole tickets |
| Handle time varies a lot by agent | Read the slow tickets (7) | Easy | All three |
The two easiest wins are the first two. Both take about a week and neither needs you to buy anything.
Join our newsletter
The latest tips and deals delivered straight to your inbox.
Four things not to do
- Giving each agent a handle time target without changing anything else. They will hit it by cutting conversations short, and fewer problems will get solved.
- Putting a timer on the agent’s screen during a ticket. It pushes them to close the ticket, not to fix the problem.
- Comparing agents who handle different kinds of tickets. That compares the queues, not the people.
- Judging any change after one week. Handle time moves around with holidays, staffing, and product releases. One week tells you nothing.
How do you know if it worked?
Watch three numbers together. Handle time on its own will mislead you.
| What to watch | What it tells you | What you want to see |
|---|---|---|
| Average handle time | What a ticket costs you | Slowly going down |
| Solved on the first try | Whether the problem actually got fixed | Holding steady or going up — if it drops, stop |
| Customer satisfaction | Whether the experience held up | Holding steady or going up |
| Your slowest one in ten tickets | Where the real pain is | Going down, and often more useful than the average |
That last row matters more than it looks. Averages hide the tickets that damage relationships with customers, and your slowest tickets are usually where the fixable problems are. If you only have time to watch one number, watch that one.
Break everything down by ticket type and by channel. One average covering phone, chat, and email answers no question you actually have. You can track it in LiveAgent , or in any tool that shows handling time per queue.
How LiveAgent helps
LiveAgent is our product; this guide covers it alongside alternatives. Everything above works in any help desk that has saved replies, a knowledge base, and routing rules. We built ours into one place because switching between tools is itself a thing that slows agents down.
Here is how the seven map onto LiveAgent. Canned messages cover the saved replies in point 2. The internal knowledge base covers point 3. Routing rules and the phone menu cover point 4. Ticket validation covers point 5, and the agent report gives you the per-queue breakdown that points 6 and 7 need, including work time, answered tickets, and average answer times.
For the wrap-up work in point 1, the AI Answer Composer drafts replies from your own content. It is available on trials and on the Small business, Medium business, Large business, and Enterprise plans, and is not included in the Legacy or Free plans. LiveAgent does not charge a setup fee for it; AI usage is billed through FlowHunt AI credits, so what you pay depends on your ticket volume and which model you choose.
Where to start this week
If you are working out how to reduce average handle time in your own team, remember that it comes down when the work around the conversation disappears, not when the conversation gets shorter. Pick two things: watch what your agents do after a ticket closes, and write saved replies for your ten most common questions. Keep “solved on the first try” on the same screen as handle time so you can see if anything starts slipping. And check your target against real industry benchmarks before you commit to a number. If you want the fuller picture, our guide to twelve help desk metrics covers what to watch alongside this one.
