Average Handle Time (AHT) Calculator
Free average handle time calculator: AHT in mm:ss from talk, hold and after-call work, plus blended vs human AHT after AI. Formula, example, benchmarks.
Use one channel at a time. Everything runs in your browser, and the inputs are saved in the page link so you can share the result.
Results
Updates live
Average handle time
10:00
10.00 min · 600 seconds per contact
Where the time goes, per contact
- Talk or chat
- 7:12 (72%)
- Hold
- 0:48 (8%)
- After-call work
- 2:00 (20%)
Total handle time 17 h over 100 contacts. Plan phone staffing with this AHT in the Erlang C calculator.
How this is calculated
In "From totals" mode it adds total talk or chat time, hold time and after-call work, divides by contacts handled and shows the result in minutes and seconds with the share of each part. In "AI vs human split" mode you enter contacts and time per contact for three groups: handled by AI only, started by AI and handed to a person, and sent straight to a person. It returns the blended AHT across all contacts, the human AHT across contacts a person touched, each group's AHT and total human handle time, and compares them with your AHT before AI if you enter it. Everything runs in your browser.
What average handle time measures
Average handle time (AHT) is the time spent on each contact from start to finish: talk or chat time, plus hold time, plus after-call work, divided by contacts handled. 1,000 minutes of handling over 100 calls is an AHT of 10:00. Enter your totals, or split contacts into AI-only, handoff and human-only groups to see what a blended number hides.
AHT feeds every staffing model, so a wrong AHT means the wrong headcount. After an AI agent goes live it also becomes easy to misread, which is why the calculator reports a human AHT next to the blend.
Average handle time formula
- AHT = (total talk time + total hold time + total after-call work) ÷ contacts handled
- Blended AHT after AI = (AI time + human time across all contacts) ÷ all contacts
- Human AHT = human talk + hold + wrap time ÷ contacts a person touched
Call Centre Helper uses the same calculation with wrap time as the third term and warns that leaving hold time out lets agents lower AHT by parking callers on hold. For email and chat, Zendesk's AHT guide gives a separate formula per channel, so calculate each channel on its own and never mix voice and email in one number.
Worked example A: AHT from totals
A voice queue handles 100 calls in a day, with 720 minutes of talk, 80 minutes of hold and 200 minutes of after-call work.
| Step | Calculation | Result |
|---|---|---|
| Total handle time | 720 + 80 + 200 | 1,000 min |
| AHT | 1,000 ÷ 100 | 10:00 (600 s) |
| After-call work share | 200 ÷ 1,000 | 20% |
To combine days or agents, add the totals and divide once. A day with 10 calls at 20 minutes and a day with 90 calls at 8 minutes give (200 + 720) ÷ 100 = 9:12, not the 14:00 you get by averaging the two daily figures.
Worked example B: AI vs human split
Before AI, a team handled 1,000 contacts a month at 10:00 each, 10,000 human minutes. After launch the same 1,000 contacts break down like this:
| Group | Contacts | Time per contact | Total |
|---|---|---|---|
| AI only | 500 | 2 min AI | 1,000 min |
| AI, then a person | 200 | 2 min AI + 14 min human | 3,200 min |
| Straight to a person | 300 | 11 min human | 3,300 min |
| All contacts | 1,000 | 7:30 blended | 7,500 min |
Blended AHT falls 25%, from 10:00 to 7:30. Human AHT rises 22%, to (2,800 + 3,300) ÷ 500 = 12:12, because AI kept the quick contacts. Human workload falls 39%, from 10,000 to 6,100 minutes.
All three numbers are true at once. Report the blend as a cost figure, staff from the human AHT, and judge agents only against contacts of the same type. The AHT after AI guide adds same-issue re-contacts, which a blended AHT never shows.
Average handle time benchmarks
Published AHT figures differ in what they count and how they were gathered. As of November 2026, these are the figures the pages below publish:
| Segment | Published figure | Source and basis |
|---|---|---|
| All industries (2021) | 424 s (7:04) | SQM Group benchmarking study; talk + wrap-up, no hold |
| General guideline | "around six minutes" | Zendesk; no study cited |
| Retail | 3–4 min | Zendesk estimate; no study cited |
| Travel and hospitality | 3–5 min | Zendesk estimate; no study cited |
| Banking and financial services | 4–6 min | Zendesk estimate; no study cited |
| Utilities | 4–6 min | Zendesk estimate; no study cited |
| Telecommunications | 5–7 min | Zendesk estimate; no study cited |
| Insurance | 5–7 min | Zendesk estimate; no study cited |
| Healthcare | 6–8 min | Zendesk estimate; no study cited |
| Technical support | 8–10 min | Zendesk estimate; no study cited |
SQM Group's operational data findings report that "The average handle time (talk time + wrap-up time) is 424 seconds" in its 2021 study. The industry ranges come from Zendesk's average handle time guide, which calls them "estimates of standard industry benchmarks".
Use them to sanity-check a number, not as a target. Your own trend for one channel and contact type, measured the same way each month, is the comparison that holds up.
Related RTCI tools and guides
- Average handle time after AI automation
The cohort model behind the split mode, with re-contacts and journey minutes.
- Erlang C Calculator
Turn the human AHT into agents on phones for a service-level target.
- Customer Support Staffing Calculator
Monthly headcount for email and ticket queues from handle time per ticket.
- Call Center Occupancy Calculator
How much of available time goes to handling contacts.
- First Contact Resolution Calculator
The same AI vs human split for FCR, so a shorter AHT is not bought with repeat contacts.
- Resolution rate vs deflection rate
Which contacts belong in the AI-only group.
Frequently asked
How do you calculate average handle time?
Add total talk time, total hold time and total after-call work, then divide by the number of contacts handled. 720 minutes of talk, 80 of hold and 200 of after-call work over 100 calls is 1,000 ÷ 100 = 10 minutes, or 10:00. To convert seconds to minutes and seconds, divide by 60: 424 seconds is 7:04.
What is a good average handle time?
There is no universal target. SQM Group's 2021 benchmarking study reported 424 seconds (talk plus wrap-up, without hold) across the call centers it measured. Zendesk publishes estimates from 3–4 minutes in retail to 8–10 minutes in technical support without citing a study. Compare against your own trend for the same channel and contact type.
Should AI-handled contacts count in AHT?
Report both ways. Blended AHT counts AI and human time across every contact and shows total cost. Human AHT counts only contacts a person touched and is the number to staff from. With 500 of 1,000 contacts handled by AI at 2 minutes, the blend was 7:30 while human AHT was 12:12.
Why did our average handle time go up after adding AI?
AI usually takes the quick, repetitive contacts, so the ones left for people are longer. In the example on this page human AHT rose from 10:00 to 12:12 (+22%) while total human handle time fell 39%. Check the workload before treating the higher AHT as a performance problem.
Can I average daily or per-agent AHT figures?
Only if you weight them by contacts. Add the handle-time totals and divide by total contacts. A day with 10 calls at 20 minutes and a day with 90 calls at 8 minutes average 9:12, not 14:00.