Count contribution and capacity separately — and plan for the downside.

Inputs you need

Use your own records for a recent, normal month. Rough numbers are fine if you note where they came from.

• Monthly requests by channel
• Share that become completed bookings or orders
• Contribution per completed booking or order
• Staff minutes spent per request today
• Monthly operation and usage cost
• One-off setup cost and the period to spread it over

Use contribution, not revenue

A completed booking worth a certain amount in revenue does not add that amount to profit. Subtract the variable cost of delivering it — materials, commissions, delivery, payment fees — and use what remains. If extra bookings need extra staff, include that too.

Treat setup separately

Spread one-off setup over the period you expect the workflow to run unchanged, often twelve months, and add it to the monthly cost. Also report the payback period on its own so the setup decision is visible.

Hypothetical worked example

Hypothetical: every number below is invented in generic currency units to show the maths. They are not HumAI prices or results. A business receives 600 requests a month and completes 50% of them (300). After launch it completes 55% (330): 30 extra bookings at a contribution of 40 units each is 1,200 units. Monthly operation and usage cost 700 units; setup of 3,600 units spread over 12 months adds 300, so the monthly cost is 1,000 units.

• Hypothetical net: 1,200 − 1,000 = 200 units per month
• Hypothetical break-even: 1,000 ÷ 40 = 25 extra bookings
• Hypothetical cost per completed booking: 1,000 ÷ 330 ≈ 3 units
• Downside (52% completion, 12 extra bookings): 480 − 1,000 = −520 units

Capacity freed is not cash saved

In the same hypothetical, saving four minutes on each of 600 requests frees 40 staff hours a month. That is valuable capacity — faster replies, more follow-up, less overtime — but it only becomes a cash saving if rotas, overtime or hiring plans actually change. Report it separately from contribution.

Check the result after launch

Agree the measures before launch and review them monthly against the baseline. If the downside case is what happens, change the scope or stop; a decision based on evidence is the point of the calculation.