Salesforce Agentforce Shifts to Pay-Per-Resolution Pricing Model
Salesforce has announced a significant change to its Agentforce pricing strategy, moving to a "pay-per-resolution" model for its Agentforce Help Agent. This new model aims to align costs directly with successful customer outcomes, rather than based on usage or user licenses.
Key Aspects of the Pay-Per-Resolution Model
The core of the new pricing is that organizations are only charged when the Agentforce Help Agent autonomously resolves a customer issue from initiation to completion. If a customer requests human intervention or expresses dissatisfaction, no charge is incurred. In such cases, the agent is designed to pass relevant context to the service team for escalation.
During these agent interactions, Data 360 (formerly Data Cloud) is unmetered. This approach is intended to eliminate the need for customers to forecast consumption or worry about overages, as the investment is directly tied to successful resolutions.
This pay-per-resolution pricing is scheduled to become available in July 2026.
Evolution of Agentforce Pricing
Agentforce's pricing has been a point of discussion since its announcement at Dreamforce '24. Initially, the messaging focused on a "$2 per conversation" model. However, this model faced criticism for lacking flexibility and obscuring total costs, particularly for interactions that did not deliver tangible value.
This earlier usage-based model was a departure from Salesforce's traditional seat-based pricing but still presented challenges. Concerns were raised about predictable ROI and the lack of guaranteed outcomes, leading to slower adoption.
In response, Salesforce introduced "pricing innovations" including Flex Credits, a consumption-based model designed to link costs to outcomes. This was further supported by Flex Agreements and new Agentforce user licenses, aimed at addressing community hesitancy around consumption-based models where outcomes were not guaranteed.
The shift from seat-based to consumption-based, and now to outcome-based pricing, indicates Salesforce's responsiveness to customer feedback and its confidence in Agentforce's capabilities.
Agentforce Help Agent Capabilities
The Agentforce Help Agent is equipped with a library of actions to manage cases, schedule appointments, and update orders across various customer touchpoints, including voice, web, portal, and messaging.
Salesforce cites its internal experience: on help.salesforce.com, Agentforce has reportedly handled 4.3 million inquiries and resolved 70% of them. Insights from this internal deployment are integrated into the current offering.
Both Agentforce Help Agent and Agentforce Customer Service Portal are slated for general availability in July 2026.
Considerations for Outcome-Based Pricing
Implementing an outcome-based pricing model presents challenges. Salesforce incurs costs related to token generation and compute power for AI models. For the pay-per-resolution model to be financially viable, token and compute costs must be low enough to ensure Salesforce's profitability, and the AI model must maintain a high success rate.
Additionally, there's a potential vulnerability to malicious actors who might repeatedly engage the agent and claim dissatisfaction to drive up costs without achieving resolution. Salesforce's adoption of this model suggests a high degree of confidence in its AI's effectiveness and a strategy to mitigate such risks.
Key Takeaways
- Salesforce is moving Agentforce AI pricing to a "pay-per-resolution" model, available July 2026.
- Customers are charged only when the Agentforce Help Agent autonomously resolves an issue.
- This is a shift from previous "per conversation" and "seat-based" pricing models.
- The new model aims to tie costs directly to successful customer outcomes.
- Challenges include managing AI model costs and mitigating potential abuse of the system.
Leave a Comment