AI changes what a software company can charge for
A working builder note on why AI agents make outcome pricing more natural than seat pricing in some workflows.
When software performs the work rather than merely helping a person perform it, the economically natural unit can shift from access to completed outcome.
Traditional software usually sells access. A human logs in, uses the tool and performs the work.
AI agents complicate that model because the software can increasingly perform the work itself.
From tool to worker
If the product schedules the interview, resolves the conflict, follows up and completes the workflow, charging per recruiter seat starts to feel disconnected from the value created.
Charging for an outcome can be cleaner. But it is also harder: now the product company owns much more of the operational risk.
That pressure is useful. It forces the system to handle exceptions, measure completion and become accountable for what actually happened.