Per-seat pricing was written for a world where marginal cost was zero. That world is over. In an AI product you do not price the seat, you price the work the seat is no longer doing.
Why per-seat breaks
Cost-to-serve is now the largest line item on my COGS, and it scales with what the user does, not whether they have a license. A power user at $50 a seat might cost me $200 a month to serve. A dormant seat costs me $50 and earns the same. I am subsidizing power users with dormant ones until the power users add more, the subsidy collapses, and gross margin disappears in a quarter. Customers figured this out faster than most vendors did: they buy fewer seats and use them harder. Seat count goes down as usage goes up. That is the death spiral. Companies still pricing per-seat in 2026 run gross margins 40 percentage points below companies on hybrid or outcome-based pricing.
The four models that work now
- Hybrid: base fee plus usage. A floor that covers fixed costs, usage on top tied to a value metric. Adoption jumped from 27 percent to 41 percent of AI software companies in a single year. Safest move off per-seat.
- Outcome-based. You charge when the agent succeeds: per resolved ticket, per successful translation, per document processed. Highest alignment, hardest to instrument, highest pricing power.
- Tiered consumption. Customers buy a bucket of units up front with overage. Predictable, easy to sell, mediocre alignment but better than seats.
- Pure usage. Pay for what you use, no minimum. Best alignment with cost, worst predictability. Only for genuine self-serve products.
Picking the value unit is the real work
Most teams pick the wrong unit and find out six months later that customers are optimizing against it. Bad units are tokens, API calls, and compute time. Good units are a successful outcome, an active workflow run where failed runs do not bill, or a document, conversation, or session. The litmus test I use: can the customer tell me, before signing, what 100 of my value units would do for them? If yes, the unit is right. If no, I am pricing something they do not understand and I will churn them on their first big bill.
You also cannot price what you do not measure. Before changing pricing I need a defined successful outcome for every billable surface, real-time tracking per customer, cost-per-outcome telemetry, a billing system that reconciles usage to invoices, and a customer-facing dashboard. That last one is the trust contract: customers accept usage pricing only if they can see usage as it happens. Build the dashboard before you change the price. For the metric behind the margin math, see Gross Margin Is Your Job Now.
Do this this week
Do not redesign your pricing page. Pick one billable surface, write down the value unit the customer actually cares about, and start instrumenting it today even if you do not bill on it yet. That single unit is where your whole pricing redesign starts.