AI agents clear the easy contacts first, so what’s left for human agents is the hard, slow work — sizing the team on your average handle time runs low every time. And the share AI can’t resolve always comes back, so there’s a floor of human agents you can’t staff below.
AI moves headcount two ways at once: checking AI’s work pushes the floor up, AI assisting your team pulls it back down. Set both below and watch where it settles.
For most teams the real win from AI isn’t a smaller team — it’s handling more contacts without hiring at the same pace. Set your expected growth and see how many hires AI absorbs.
Live result - recalculates as you change the numbers below
Human agents you still need
human agents
Cost per contact
AI fees + human-agent time, across every contact
Weekly team cost
human agents × all-in weekly cost each
How the human-agent count builds up, at % routed to AI
Start from the count a flat average gives you, then each step is what reality adds or removes. The last box is your real number.
Your numbers
How many human agents
What AI does to that
Volume
AI agents
Human agents
AI cost
Oversight
human-agent minutes per AI resolution
AI assist
e.g. AI assists 40% of contacts and makes each 30% faster = 12% saved
Growth
% more contacts per week
Human agents you still need, as AI agents take more
Blue = what you really need. Grey = what a flat average says. Red = the floor, even at 100% AI. The blue–grey gap is your shortfall.
Cost per contact (AI fees + human-agent time)
Why there's no "best" coverage: in this model, cost only ever bottoms out at one of the ends — all AI or none, never the middle. The badge above names which end wins for your inputs.
Clean-math version. Real life can bring a middle sweet spot back: whole-person hiring, minimum shift staffing, bulk AI discounts, flat platform fees.