How to calculate the number of agents?

The tool converts volume and AHT into work hours, adjusts concurrency and peak, and divides by the useful capacity of each agent after occupancy and unavailability.

Does the sizer use Erlang C?

No. It performs a monthly capacity sizing. Voice operations that need to guarantee probability of answering within seconds should complement the analysis with arrivals per interval and a queueing model such as Erlang C.

It is the average number of conversations a person handles at the same time. For voice, it is typically one. For digital channels, the value depends on complexity and process.

Does the result replace a schedule by time slot?

No. The result estimates monthly capacity. Formal schedules require intraday distribution, breaks, legislation, skills, absences, and service levels per queue.

groups Service capacity

How many people does your operation need to serve without exceeding capacity?

Convert interactions and average handle time into workload. Adjust channel, concurrency, occupancy, unavailability, and peaks to compare current team with estimated need.

visibility Open resultcalculate Formula explaineddevices Voice and digital channels

Volume, shift, and productivity

Describe monthly capacity

Start with the main channel or use weighted averages from a queue with similar behavior.

support_agent

Main channel The selection suggests concurrency; you can change the value.

Interactions received per month

contacts

Current agents

people

Average handle time per interaction

min

Average active handling time.

Simultaneous conversations

per person

Operating days per month

days

Scheduled hours per day

Desired productive occupancy

Portion of available time occupied by interactions.

Breaks and unavailability

Training, meetings, breaks, absences, and auxiliary tasks.

Peak protection factor

1.25 adds 25% protection over average load.

First response target

min

Registered in the diagnosis; does not replace an intraday queue model.

infoThis calculation sizes monthly capacity. For interval scheduling and formal service level, use arrival data every 15 or 30 minutes and a queue model.

groups

The calculated capacity will appear here

You will see headcount needed, difference from current team, productive load, and estimated capacity.

check_circle Required teamcheck_circle Shortfall or buffercheck_circle Monthly capacity

Monthly sizing

Suggested next actions

How we arrived at the number

chat WhatsApp work LinkedIn

Receive capacity analysis

Share the scenario to review queues, schedules, channels, automations, and overflows.

Notice: The result is a monthly capacity estimate, not a work schedule or SLA guarantee. Intraday volumes, skills, legislation, actual breaks, absenteeism, and queue distribution must be validated in the project.

architecture Capacity model

The calculation separates load, availability, and peak protection

This allows you to understand which assumption is driving headcount needs and simulate operational changes without hiding the formula.

Convert volume into hours

Interactions × AHT, divided by concurrency, yields the monthly active load.

Adjust real capacity

The shift is reduced by breaks and unavailability, and then by the desired productive occupancy.

Protect against peaks

The peak factor adds buffer to average volume. An hourly analysis is still needed for formal scheduling.

Questions about sizing

What concurrency should I use?expand_more

Use 1 for voice. For messaging, measure how many conversations an agent can maintain with quality. The more complex or sensitive the interaction, the lower the sustainable concurrency tends to be.

What counts as unavailability?expand_more

Breaks, meetings, training, absences, feedback, and tasks that remove the person from the queue. Use actual schedule data whenever available.

Why limit occupancy?expand_more

Planning 100% of available time as service leaves no room for variation, after-call work, transition, and waiting. The appropriate percentage depends on the channel and service level.

When to use Erlang C?expand_more

For voice or real-time queues when the operation needs to estimate waiting probability per interval. You will need to provide arrivals and AHT in smaller blocks, plus the service level target.