Agentforce Implements New Business Model to Drive Growth and Innovation

Pricing Changes: Conversation-Based Fees Evolve to Action-Based Credits

Until now, Agentforce’s business model primarily charged users €2 per conversation, establishing a straightforward billing system that revolved around the number of interactions users had with the platform. This per-conversation approach was the foundation of the pricing structure, providing a simple and predictable cost model for clients.

With the upcoming rollout of Salesforce Summer ’25, a new billing system is being introduced: the “Flex Credits.” This new model moves away from counting conversations and shifts toward measuring “actions.” In this context, an “action” refers to any specific function performed by an agent on the platform. For example, standard functions include generating email drafts, scheduling appointments, enhancing product descriptions, sending meeting requests, or extracting specific values from user inputs.

Currently, the same pricing applies to both standard features and customized functions. Each action, whether in a live environment (production) or a testing sandbox, costs 20 and 16 credits respectively. This uniformity simplifies the billing, but also signals a transition toward more granular, action-based billing.

Converting Licenses into Agentforce Credits: A New Option

Salesforce is offering options for users to convert existing licenses into Flex Credits. Some packages include bundled credits, which can be purchased in packs of 100,000 for €500—essentially making each action worth about €0.10. This approach provides a more flexible and granular way to manage costs, especially for internal agents who use the platform regularly.

This system offers a more detailed and specific billing approach compared to the conversation-based model, making it better suited for organizations that need precise control over their agent usage. To facilitate this, Salesforce has introduced new agreements called “Flex Agreements,” designed to allow organizations to convert seats (licenses) into credits and vice versa, providing flexibility depending on operational needs.

Further, Salesforce plans to release agent licenses by user, targeted for this summer across various verticals, including Sales, Service, Field Service, and Industries. These licenses will be designed to integrate seamlessly with the credit system, integrating internal agent management with the new billing method.

The original conversation-based approach remains in place for now, but the two systems cannot be used simultaneously within the same Salesforce organization. Organizations must choose between all conversation-based SKUs or all credit-based SKUs for consistency. Combining both billing models within a single organization is not currently supported.

Additionally, once an subscription expires, any remaining credits are non-transferable, emphasizing the importance of planning and accurate usage. Each action also consumes a maximum of 10 tokens, a unit for processing language tasks, which Salesforce tracks strictly. If the number of tokens used exceeds this cap, additional tokens are deducted accordingly. It is important to note that this credit consumption does not include any supplementary rights, such as Einstein for Sales or Einstein for Services add-ons, nor credits for actions built with Apex, like Einstein Request.

Additional Resources for Context and Future Developments

– The evolving branding of Salesforce, shifting away from concepts like copilots and chatbots toward empowering human agents
– The role of agent AI in unified communication platforms
– A taxonomy of AI agent protocols, extending beyond the MCP framework
– From traditional UX to AX (Agent Experience), highlighting the need to design interfaces specifically for AI agents
– Ethical and strategic concerns, exemplified by companies like Klarna, questioning the limits of generative AI deployment

Dawn Liphardt

Dawn Liphardt

I'm Dawn Liphardt, the founder and lead writer of this publication. With a background in philosophy and a deep interest in the social impact of technology, I started this platform to explore how innovation shapes — and sometimes disrupts — the world we live in. My work focuses on critical, human-centered storytelling at the frontier of artificial intelligence and emerging tech.