Most software is sold by the seat, and a seat is a person. AI breaks the link between the work a customer gets done and the number of people doing it, which is good news for the customer and bad news for a price that counts people. The companies in the atlas are answering in different ways: Microsoft sells an AI tier above the seat, Snowflake sells by consumption, and ExlService is moving its contracts toward transactions and outcomes. Before you pick a direction, do the arithmetic for your own business.
The repricing calculator
Enter what you charge per seat today, how many seats AI could cost you, and what a price per task or per result would look like. Every figure is a month. Change any input and the results follow.
Paying customers today.
200The average number of paid seats or users per customer.
50The share of seats your customers could stop paying for once AI takes on part of the work. Pick a number you could defend to your board.
30%The unit you would charge for instead: tickets resolved, documents drafted, reports run, calls answered.
120Work that gets cheaper usually gets done more. How much more volume would customers send?
25%What you would charge for each task under usage or outcome pricing.
$0.4Your cost to run each task: computing power, model fees and any review by people.
$0.05Seat revenue at risk: $120,000 · Price per task that matches today's seat revenue: $0.27
Results
| Value | |
|---|---|
| Seat revenue today | $400,000 |
| Seat revenue at risk | $120,000 |
| Seat revenue after the cut | $280,000 |
| Tasks a month under the new price | 1,500,000 |
| Revenue at your price per task | $600,000 |
| Gross profit per month at that price | $525,000 |
| Task revenue against today's seat revenue | 150% |
| Price per task that matches today's seat revenue | $0.27 |
| Gross margin per task | 87.5% |
Your inputs
| Value | |
|---|---|
| Customers | 200 |
| Seats per customer | 50 |
| Price per seat, a month | $40 |
| Seats customers could drop as AI does the work | 30% |
| Tasks per seat, a month | 120 |
| Extra tasks once AI is doing them | 25% |
| Price per task | $0.4 |
| Your cost per task | $0.05 |
About the method
- Seat revenue at risk assumes the price per seat does not change; it is the revenue the seat model loses if customers need fewer seats.
- Volume under the new price is today's seats times tasks per seat, raised by the extra work customers send once it is cheaper.
- The break-even price per task is today's seat revenue divided by that volume. Charge more and the switch earns more than seats do today; charge less and it earns less, before any change in the number of customers.
- Gross profit counts only your cost per task. It leaves out the fixed costs of building and selling the product.
A thinking aid from Stratrix. The arithmetic is exact; the inputs are your assumptions, and the result is only as good as they are. Nothing here is a forecast or financial advice.
Get the full report: PDF report, Word (editable), Excel
How to read the result
- Seat revenue at risk
- What the seat model loses if customers drop the seats you entered. It is the size of the problem, not a forecast.
- The break-even price per task
- The price at which charging per task brings in what seats bring in today. If your customers would not pay that much for each task, a straight switch costs you revenue, and a hybrid of seats and tasks may be the better answer.
- Gross margin per task
- A price per task has a cost per task behind it. If computing power or review by people takes most of the price, the volume that makes revenue grow can also make losses grow.
Questions
- Should AI products be priced per seat or per task?
- It depends on whether AI shrinks the number of people who use the product. If customers need fewer seats once AI does the work, a price per seat shrinks with them, and a price per task, per unit of use or per result keeps revenue tied to the work done. Many companies sell both: a seat for access and a meter for AI work.
- How do I find the break-even price for usage-based pricing?
- Divide today's seat revenue by the number of tasks customers would run under the new price, including any extra volume once the work is cheaper. The result is the price per task at which revenue stays the same.
- What is outcome-based pricing?
- The customer pays per result, such as a resolved support ticket or a processed claim, rather than per seat or per hour. It ties the seller's revenue to what the customer values, and it moves the risk of delivering the result to the seller.
More from Stratrix
“Where does the money actually come from, and who sets the rate it comes in at?”
See the pack — $499 →The Business Model Wizard
Make the seven choices that define your model, test your numbers against real companies, and meet the rival AI would build against you.
Free · about twelve minutesBuild your Model Thesis →Test a decision, free
Set a live decision against the companies that made it before you.
AcquisitionsThe Distance TestDivestituresThe Divestiture TestDiversificationThe Diversification TestMake or buyThe Make-or-Buy TestWeekly email
Stratonomics, every Tuesday
One data story every Tuesday: what happened, what it means, and the figures you can quote. Free.