AI Pricing Shifts to Outcome-Based: Salesforce's Pay-Per-Resolution Agent, Bairong Yunchuang's Result Cloud

The pricing model for AI services is shifting from billing based on token consumption to outcome-based payment. Outcome-based payment means settling charges based on the actual business problems the AI solves and the quantifiable value it creates, rather than on token consumption or the number of calls.

Some companies have already put this into practice. In June 2025, software service provider Salesforce launched its Agentforce Help Agent, which adopts a pay-per-resolution model. The system only charges when the agent autonomously resolves a problem from start to finish; if a customer provides negative feedback or requests a transfer to a human, no fee is incurred. In December 2025, Bairong Yunchuang released its Result Cloud Platform, which supports three value-exchange models: per-task pricing, per-job compensation, and value-based revenue sharing.

Currently, token-based billing is the mainstream method, offering an objective, auditable, and uncontested basis for measurement. However, outcome-based payment still faces multiple obstacles in moving from concept to large-scale implementation. There is no consensus on the definition of outcomes and measurement standards; billing boundaries are ambiguous, as business results are often the product of multiple interconnected steps, making it difficult to attribute them solely to a single model call; and settlement cycles are mismatched, since token billing is instantaneous while outcome-based payment requires post-event verification of results.