Results are announced here.
Your results appear here.
Estimates for planning only. Not financial, legal or staffing advice.
How this call center staffing calculator works
Enter how many calls arrive in an interval, how long each one takes and the service level you promise. The calculator finds the smallest number of agents that meets the target with the Erlang C queueing model. Then it repeats the calculation for the calls that stay with people after an AI voice agent takes its share.
The Erlang C formula
Traffic in Erlangs is calls × average handle time ÷ interval length in seconds. For N agents and traffic A, the probability that a call has to wait is:
Pw = [AN ÷ N! × N ÷ (N − A)] ÷ [Σk=0…N−1 Ak ÷ k! + AN ÷ N! × N ÷ (N − A)]
Service level is 1 − Pw × e−(N − A) × target time ÷ AHT, and the average speed of answer is Pw × AHT ÷ (N − A). The calculator adds one agent at a time until the service level reaches your target. It evaluates the formula through the Erlang B recursion, which stays accurate for large contact centers where factorials overflow.
Example: 100 calls in 30 minutes with a 180-second AHT is 10 Erlangs of traffic. 13 agents answer 79.6% of calls within 20 seconds and 14 agents answer 88.8%, so an 80/20 target needs 14 agents on the phones.
Shrinkage and occupancy
Shrinkage is the share of paid time when agents cannot take calls: breaks, training, meetings, absence. Agents to schedule = agents needed ÷ (1 − shrinkage). Occupancy is the share of logged-in time agents spend handling calls: traffic ÷ agents on the phones. The table under the result shows how service level, waiting time and occupancy change with a couple of agents more or fewer.
Where an AI voice agent fits
An AI voice agent can complete routine calls on its own, such as payment-date questions, order status, appointment changes and simple account questions. Every call it completes leaves your queue, so fewer people are needed for the same service level. The calculator assumes the calls the AI agent takes are as long as the average call: if they are shorter, the calls left for people are longer on average and the saving is smaller. The share you enter is your estimate, so check it on a pilot before you change the roster.
Frequently asked questions
What is the formula for calculating call center staffing requirements?
First convert calls into traffic: calls per interval × average handle time in seconds ÷ seconds in the interval. That number of Erlangs is the minimum workload. Erlang C then finds how many agents you need above it to answer a target share of calls within a target time. Dividing by (1 − shrinkage) gives the number of agents to schedule.
What is the difference between Erlang B, Erlang C and Erlang A?
Erlang B assumes a call that finds every line busy is lost, so it suits trunk and line sizing. Erlang C assumes every caller waits in the queue until answered, which suits inbound contact centers. Erlang A also models callers who hang up while waiting, so it needs one more input: how patient callers are.
What is an Erlang?
One Erlang is one hour of call time per hour: one line or one agent busy all the time. For example, 100 calls of 3 minutes in an hour add up to 300 minutes of calls, or 5 Erlangs.
What does an 80/20 service level mean?
It means 80% of calls are answered within 20 seconds. Both numbers matter: a stricter answer time or a higher share needs more agents, and the extra agents needed grow quickly as the target gets close to 100%.
Why do I need more agents than the traffic in Erlangs?
With exactly as many agents as Erlangs, every agent is busy all the time and the queue keeps growing. The extra agents absorb random bursts of calls. How many you need depends on your service level target, and the share of extra agents falls as the contact center gets bigger.
Does Erlang C account for callers who hang up?
No. Erlang C assumes nobody abandons the queue and that calls arrive at random at a steady average rate through the interval. Callers who hang up clear the queue faster, but those who call again add calls, and a busier stretch inside the interval needs extra agents. Count repeat attempts in your call volume and check your busiest minutes separately.
How does an AI agent change the number of agents I need?
Calls the AI agent completes never reach your queue, so traffic drops by the same share. Queues are not linear, so the number of agents falls by a different amount than the share of calls. The calculator shows both scenarios side by side.