Strongly agree. While token prices may fall, we’re also asking AI to do far more than we were a year ago, larger contexts, more reasoning and increasingly agentic processes. That can offset the savings and make costs harder to predict. For many businesses, tying pricing to business outcomes rather than token consumption feels like a much clearer way to align cost with value.
I’d be in favor of outcome-based pricing for enterprise AI, because token pricing is a nice engineering metric but a bad business contract. Tokens measure consumption. Outcomes measure value and enterprises should pay for value, not for how chatty the model happened to be.
Outcome based pricing is a better fit when the buyer cares about business value more than raw model usage. Token pricing is simpler to start with, but it often pushes cost risk onto the customer and makes budgeting harder as prompts, context, and agent behaviour grow.
Outcome-based pricing aligns spend to something executives actually care about: resolved tickets, completed workflows, qualified leads, or other measurable results. That makes procurement and finance conversations easier because you are paying for value delivered, not for hidden consumption behind the scenes. For enterprise AI, that predictability is a big deal. It reduces “surprise bills,” makes ROI reporting cleaner, and encourages teams to use AI where it truly moves the needle rather than worrying about every extra token.
Token-based pricing is still useful when you are experimenting, workloads vary a lot or you want a direct meter for technical consumption. It is also transparent in a technical sense: you can see exactly what you are paying for. The downside is that the meter tracks computation, not business value, so cost can rise even when the outcome is mediocre.
Outcome-based pricing is not free of complexity. You have to define outcomes carefully, avoid gaming the metric, and agree on attribution when multiple systems contribute to the result. So it is stronger commercially, but harder operationally
Outcome-based pricing makes far more sense for enterprise AI. Business leaders want to understand the cost of achieving a result—not track tokens, context windows, or model calls.
The key will be keeping outcomes clearly defined and transparent. If done well, this can make AI adoption easier to justify, forecast, and scale.
Outcome-based pricing aligned to business cases