How many agents do you need per 100 inbound calls?

For 100 inbound insurance calls per hour, plan on roughly 5 to 8 licensed agents as a starting point, assuming average handle times of 8-10 minutes and a service level target around 80/20. That number moves fast once you factor in shrinkage, occupancy, and whether you're running Medicare during AEP or a steady auto insurance book in June.

Here's the thing: everyone wants a single number. Agencies call me asking "just tell me the ratio," like there's one answer that works for final expense in February and Medicare Advantage in November. There isn't. But there is a real formula. Once you understand it, you can build your own staffing model instead of guessing and hoping your abandonment rate doesn't blow up your carrier relationships.

The math behind the ratio

Start with Erlang C, the staffing model most call centers still lean on, referenced constantly by groups like ICMI. In practice, Erlang C tells you roughly 1 agent per 12-20 calls per hour, depending on how complex the calls are and what service level you're targeting. That's a wide range on purpose. A simple auto quote call running 6 minutes behaves nothing like a Medicare Advantage call running 12-15 minutes with compliance scripting baked into every sentence.

Do the basic math first. If your average handle time is 9 minutes, one agent can theoretically handle about 6.6 calls per hour if they did nothing else. For 100 calls per hour, that's roughly 15 agents in a pure "everyone is always on the phone" world. Nobody lives in that world.

That's where shrinkage comes in. Insurance call centers typically see shrinkage rates of 25-35%. That covers breaks, training, after-call work, coaching sessions, and the fifteen other things pulling an agent off the phone even when they're clocked in. If you need 15 agents handling calls at any given moment and your shrinkage sits at 30%, you actually need to staff around 21-22 people to cover that.

Then there's occupancy. Sustainable occupancy [for insurance agents](/sales-team-training/how-to-write-a-call-qualifying-script-for/) runs 75-85%. Push your team past 90% and you'll see it in the error rate within two weeks, especially on compliance-heavy products like Medicare and final expense, where a missed disclosure isn't just bad service. It's a regulatory problem. I've watched agencies chase efficiency by pushing occupancy to 92-95% and then wonder why their NAIC complaint numbers ticked up the same quarter. Not a coincidence.

So the real formula looks something like this: raw agent need, based on AHT and call volume, divided by your target occupancy, then grossed up for shrinkage. Run that math for 100 calls per hour at 9-minute AHT, 80% occupancy, and 30% shrinkage, and you land closer to 6-7 agents actively available. Which means staffing around 8-9 total headcount to make sure that many are actually on the phone at any moment.

Why service level changes everything

This is the part agencies miss constantly. Two centers with identical call volume and identical AHT can need wildly different headcounts, purely because of the service level target they've set. A center targeting 90% of calls answered within 20 seconds (the classic "90/20" standard) can need 15-20% more agents than a center targeting a looser 70/30 standard, even with the exact same 100 calls per hour coming in.

This matters because service level isn't just a vanity metric. If you're buying inbound calls and routing them to agents, a slow answer time kills conversion. Callers hang up. Abandonment climbs. And if you're paying per call, an abandoned call because nobody picked up in time is money burned for nothing. I've sat with agencies swearing their lead quality was the problem, when the real issue was calls ringing 45 seconds before anyone picked up, and half the caller pool bailing before they ever reached a human.

If you're serious about tightening this up, decide your service level target before you build your staffing model, not after. Chasing a stricter service level after the fact means hiring in a panic, and panic hiring in insurance means compliance risk. You can't just throw a general CSR on the phone.

The licensing wrinkle nobody outside insurance deals with

Most industries can flex staffing with temp workers or newly trained reps. Insurance can't, not easily anyway. NAIC guidance and state-level rules generally require a licensed agent, not just any customer service rep, to handle calls involving plan recommendations or product comparisons. Your staffing pool is smaller than a typical call center's, and your cost per agent runs higher too.

Expect to pay $18-$28 an hour in wages alone for a licensed insurance agent in a call center setting, before benefits and overhead. States with tighter licensing requirements, or markets with heavy bilingual demand, push that number toward the top of the range or past it. This is one more reason "just hire more people" falls apart in practice. You can't post a job and fill seats by Friday. Licensing takes time, appointments with carriers take time, and none of that moves at the speed call volume spikes move.

AEP is a different animal entirely

Medicare Annual Enrollment Period runs October 15 through December 7 every year, and it is not a normal staffing problem. Carriers like Humana and UnitedHealthcare start planning AEP staffing months ahead because call volume during that window can run 2-3 times normal levels. If your steady-state ratio is 8 agents per 100 calls per hour, AEP might mean needing the equivalent of 16-24 agents worth of capacity for that same volume, because volume itself spikes well past 100 calls per hour for six or seven straight weeks.

Agencies that treat AEP staffing like a normal month get burned every single year. I've seen it happen in real time: phone queues stacking up, abandonment rates climbing past 15-20%, licensed agents burning out by week three because nobody planned for the sustained volume. Build your AEP staffing plan by August. Not October 10th.

If you're generating your own inbound call volume rather than buying it, staffing is only half the equation. The other half is making sure the calls you're generating actually match what your team can handle. I go through all of it, funnel design, call routing, compliance guardrails, in my book The Pay Per Call Revolution, along with a companion workbook that walks you through building the model step by step.

FAQ

Does this ratio change for auto insurance versus Medicare calls? Yes, significantly. Auto calls run shorter, often 6-8 minutes, and carry less compliance overhead, so you can staff closer to the 12-16 calls-per-hour end of the Erlang C range. Medicare and final expense calls run 10-15 minutes with heavier scripting requirements, pushing you toward 8-12 calls per agent per hour.

How do I know if my occupancy target is too aggressive? Watch error rates and agent turnover over a 60-90 day window. If occupancy sits above 88-90% and you're seeing compliance mistakes or agents quitting faster than usual, pull it back toward 80%.

Should I staff based on average call volume or peak call volume? Peak, always. Staffing to average volume guarantees you're understaffed during your busiest hours, exactly when service level and abandonment rate matter most.

Can non-licensed staff handle any part of these calls? In many states, non-licensed staff can handle basic intake, verification, or scheduling. But the moment the conversation touches plan recommendations or comparisons, a licensed agent has to take over. Check your specific state rules before building a hybrid model.