Vendors price AI in several ways: per user, per token or request, per resolved task, or on shared results. Each shifts risk differently.
Compare on the same workload
- Per seat — predictable; poor fit if only some users get value.
- Usage-based — scales with demand; needs budgets and alerts.
- Per outcome — aligns cost with results; needs a precise definition of the outcome.
- Hybrid — a platform fee plus usage.
Questions to ask
- What counts as an outcome and who judges it?
- What happens to price at ten times the volume?
- What are the exit and data-return terms?
Model the full cost at your expected volume, including the people who review the output.