Customer Support Staffing Calculator
How many support agents do you need? Calculate headcount from ticket volume, handle time and utilization, with or without AI. Formula and example.
Inputs
Monthly support tickets
Directional model only. Doesn't account for shift coverage, peak-hour staffing, or attrition buffer.
Results
Updates live
Agents needed
17
16.7 exact, rounded up
Monthly staffing cost
$76,500
$7.65 per human-handled ticket
Headcount fit
16.7 exact agents rounds up to 17: 98% of that rounded team's capacity is actually needed.
10,000 tickets/mo reach a human · 2000 total handle hours · 120 available hrs/agent
Staffing live phone or chat queues to a service level? Use the Erlang C calculator.
How this is calculated
Removes the share of tickets AI resolves before reaching a human, converts the remaining volume into total handle hours, then divides by each agent's available hours (working hours × target utilization) to get a headcount, rounded up.
How many support agents do you need?
For email, ticket and other asynchronous channels, headcount comes down to workload divided by capacity. Agents = (tickets reaching humans × handle time in hours) ÷ (working hours per agent × utilization), rounded up. Tickets reaching humans is your total volume minus whatever an AI agent resolves on its own.
This works because a ticket can wait in a queue for a few hours without harm, so you only need enough total hours across the month. Live phone and chat are different: callers will not wait, so you need enough people at the busiest half hour. For those channels use the Erlang C calculator, which sizes staff per interval against a service-level target.
Worked example: 10,000 tickets a month
- 10,000 tickets × 12 minutes = 120,000 minutes = 2,000 handle hours
- Each agent works 160 hours at 75% utilization = 120 productive hours
- 2,000 ÷ 120 = 16.7 → 17 agents; at $4,500 a month fully loaded, $76,500 a month, or $7.65 per ticket
- With AI resolving 40% first: 6,000 tickets × 12 minutes = 1,200 hours ÷ 120 = 10 agents, $45,000 a month
The 40% AI share removes 7 roles here, not 40% of 17, because headcount rounds up and the remaining tickets are often the harder ones. If escalated tickets take longer than your current average, raise the handle time when you model the AI scenario.
Choosing each input
- Tickets per month: use the last three months' average of tickets that needed a human reply. Leave out spam, auto-closed tickets and duplicates merged before anyone worked them.
- Handle time: total time an agent spends on a ticket across every reply, not just the first one. Most help desks report it as handle time or time spent per ticket. Read our average handle time guide for what moves it, or work it out from your totals with the AHT calculator.
- Working hours: 160 hours assumes a 40-hour week over four weeks. If you budget holidays, sick days and training separately, lower it (for example to 140) instead of lowering utilization.
- Utilization: the share of paid working time spent on tickets. Meetings, coaching, breaks and documentation take the rest. Planning at 100% builds a team that falls behind every busy week. Not the same as occupancy; see the occupancy calculator.
- Loaded cost: salary plus benefits, payroll taxes, software seats and a share of management. This is the number to use in the AI customer service ROI calculator as your human cost per ticket.
When the calculator's answer is too low
- You promise coverage hours. A 24/7 inbox needs people on every shift even when volume is low, so minimum staffing per shift can exceed the workload number.
- Volume is spiky. Launches, outages and holiday peaks need capacity the monthly average hides. Size for your busiest normal month, then plan overtime or contractors for true spikes.
- You are growing. Hiring and training take one to three months, so staff for the volume you expect when new hires are productive. The ticket volume forecast calculator projects that number.
- Agents specialize. If billing, technical and VIP queues are staffed separately, run the calculation once per queue and add the results. Each queue rounds up on its own.
Frequently asked
Why round the agent count up instead of down?
Because you can't hire 4.2 agents, and rounding down would understaff the team for the volume you specified.
What utilization rate should I assume?
70–85% is a common range once you account for breaks, training, coaching, and non-ticket work. 100% utilization isn't realistic for any support team.
How do I model a partial AI rollout?
Set "Share resolved by AI" above 0%. The calculator only staffs for the tickets that still reach a human.
How many tickets can one support agent handle per month?
Divide productive hours by handle time. An agent working 160 hours at 75% utilization has 120 productive hours; at 12 minutes per ticket that is 600 tickets a month. At 20 minutes it drops to 360.
Should I use this or the Erlang C calculator?
Use this for email, ticketing and other channels where a customer can wait hours for a reply. Use the Erlang C calculator for phone and live chat, where you need enough people answering in each half hour to meet a service level.