Service capacity is bought in whole people. Volume does not arrive in whole people.
That mismatch is the entire problem. An agency hires for a level of demand, then demand moves: a renewal quarter, an acquisition, a producer who lands three large commercial accounts in a month. The cost stays flat while the workload does not, and the agency is either overstaffed in February or underwater in October. Usually both, in the same year.
This piece is about changing the shape of that cost rather than the size of it.
Fixed Capacity Is a Bet on a Flat Year
Every service hire is a forecast. You are predicting that a certain volume will exist, continuously, for as long as that person is employed.
Insurance service volume is not flat. It clusters around renewal cycles, it spikes after a deal closes, and it moves when a carrier changes a process. Hiring to the peak means carrying that cost through every trough. Hiring to the average means the peak is absorbed by overtime, slipped turnaround, and the goodwill of people who will eventually leave.
Neither is a mistake anyone made. It is the only choice available when capacity comes in whole-person units.
What Variable Actually Means Here
It does not mean contractors, and it does not mean replacing your team.
It means that a portion of the work, the routine high-volume portion, is handled by capacity that scales with the volume rather than with your payroll. Your own staff stays the default. This is extra capacity when you need it. What changes is what happens above the line your team can comfortably carry.
Three things make that possible, and none of them is a staffing decision.
- The work has to be defined as tasks, not as jobs. You cannot flex a role. You can flex a queue of discrete, specified tasks.
- Routing has to be automatic. Each task goes to whatever is appropriate: a licensed person, an unlicensed person, or AI. Decided by what the task requires, not by who owns the account.
- It has to be measured. Otherwise variable just means unpredictable, which is worse than fixed.
Where the Variable Portion Comes From
Most of the flex is not human at all, which is the part operators tend to underestimate.
Across the agencies we operate, AI triage filters roughly 68% of inbound as noise before a person sees it, classifying each item in under 10 seconds at a 98% triage success rate. Roughly 44% of completed service tasks are now executed by AI. The platform routes more than 40,000 tasks a week with no manual intervention.
Capacity that behaves like software scales with volume by default. It does not need to be hired in December and justified in March.
The remainder flexes through the Service Network: licensed capacity, on demand, routed through the same system, licensed in all 50 states.
What Changes on the P&L
The absolute number matters less than its shape.
A fixed service cost is a floor. It sets a minimum you pay regardless of what the book does, and it only moves through a hiring or firing decision, both of which are slow and expensive and visible to your team.
A variable service cost tracks the work. When volume rises, cost rises with it and the turnaround does not slip. When volume falls, the cost falls too, without anyone losing a job.
That is a margin argument rather than a cost-cutting one. Cost per task is measured, and it comes down. Across the agencies we operate it has been cut roughly in half since launch. We do not publish that figure, because our internal cost is not a price, and the only number that matters to you is the one from your own book.
S&G Mitchell went from 17.9% to 60%+ EBITDA in 12 months. Same agency, same customers, same book of business. What changed was the operating model underneath.
COVU OS runs on top of your AMS, not instead of it, with 8 AMS integrations live.
Where to Start
Do not start with a staffing plan. Start with one workflow that spikes: certificates, or renewal prep in your heaviest month.
Define the tasks in it. Route them. Then look at what happened to turnaround during the spike, and whether anyone worked a weekend. That tells you more about whether the cost structure changed than any forecast will.
FLEX CAPACITY WITH THE VOLUME
Take one workflow that spikes and see what happens to turnaround when the capacity moves with it.
Frequently Asked Questions
What does it mean to make service a variable cost?
It means the routine, high-volume portion of service work is handled by capacity that scales with volume rather than with headcount. Your own team stays the default. What flexes is the work above the line they can comfortably carry.
Does this mean replacing our CSRs?
No. Your staff stays the default, and this is extra capacity for when volume exceeds it. The work that moves is the routine, rule-based kind that was crowding out everything else.
Why is fixed service capacity a problem?
Because capacity is bought in whole people and volume does not arrive in whole people. Hiring to the peak means paying for it through every trough. Hiring to the average means the peak is absorbed by overtime and slipped turnaround.
How does AI fit into variable capacity?
It is most of it. Capacity that behaves like software scales with volume by default. Across the agencies we operate, AI triage removes roughly 68% of inbound before a person sees it, and roughly 44% of completed service tasks are executed by AI.
How do we know the variable model is actually cheaper?
Measure cost per task on your own book before and after. That is the comparable number, and it is the one your own system should produce rather than a vendor.