Risk-Rated Capacity Planner

Attach probability to your staffing plans — see the odds of hitting SLA at each staffing level with your volume uncertainty. Replaces single-point capacity numbers with risk-rated recommendations.

This tool extends deterministic Erlang C staffing with Monte Carlo simulation to quantify the probability of achieving service level targets under forecast uncertainty. The Erlang C formula calculates the probability of queuing: P(wait) = [A^N/N! x N/(N-A)] / [sum(A^k/k!, k=0..N-1) + A^N/N! x N/(N-A)], where A is traffic intensity (volume x AHT / interval_seconds) and N is agent count. Service level is then: SL = 1 - P(wait) x exp(-(N-A) x target_answer_time / AHT).

The simulation runs 500 iterations per staffing level, each sampling a random volume from a normal distribution with the configured coefficient of variation (CV). This produces a distribution of achieved service levels for each headcount scenario, reported as P10 (worst case), P50 (median), and P90 (best case). The primary output is "probability of hitting SLA" — the percentage of simulations where service level meets or exceeds the target.

The tool identifies two staffing levels: the Erlang-calculated minimum (sufficient only at exact forecast volume) and the risk-adjusted recommendation (sufficient 90% of the time given forecast uncertainty). The difference is the buffer needed to absorb volume variability. Practitioners should calibrate the Volume CV parameter from historical forecast accuracy — if your WAPE is 12%, use 12% as CV. Higher CV requires more buffer agents. Operations that staff only to the Erlang number will miss SLA roughly half the time, because volume is equally likely to be above or below forecast.