Salesforce is preparing to introduce a new billing model for agentic access to its platform, with successful MCP and direct API calls made by registered agents set to consume Flex Credits.
The change forms part of Salesforce’s wider Headless Toolkit strategy and will sit alongside a new Agentic Identity framework, giving AI agents their own credentials and permissions rather than having them operate under a user’s identity. Salesforce has not yet revealed the Flex Credit multiplier for these interactions, but says customers will receive 30 days’ notice before metering begins.
What Is Changing?
At the centre of the change is Agentic Identity, which will require AI agents interacting with Salesforce to be registered directly on the platform.
Rather than operating under the identity of the user they are assisting, each agent will receive its own credentials and permissions, giving admins more control over what it can access and do.
Once registered, usage will be tracked through Digital Wallet. Every successful call made by that agent to Salesforce – whether through MCP or directly through an API – will be counted as a “Headless Platform Interaction” and consume Flex Credits. This means that the new model applies to direct API access where the traffic is classed as agentic.
Salesforce has not yet published the multiplier that will determine how many Flex Credits each interaction consumes, meaning customers cannot yet calculate the cost impact.
Agentic usage is also not currently being metered, and Salesforce says it will provide 30 days’ notice before billing begins.
The company has also stressed that traditional integrations are not being moved onto the same model. Existing API-based integrations will continue to use their current pricing and security arrangements, while the new consumption model is specifically tied to registered agentic workflows. Metering is only done in “active production” orgs, which excludes sandboxes, scratch orgs, and Developer Edition orgs. This means agentic development will not incur MCP usage costs, either.
For MCP clients and agents using APIs, customers will eventually need to reconfigure their connections to use OAuth credentials associated with the newly registered Agentic Identity.
SF Ben has reached out to Salesforce for comment.
What Does This Mean for Customers?
The security side of the change is likely to be easier to welcome than the new consumption pricing model. SF Ben Technical Content Writer Tim Combridge said the Agentic Identity framework is “a really good step in the right direction,” especially because agents will have their own identities and defined permissions rather than relying on those of individual users.
The bigger potential problem is cost. Customers bringing external AI tools such as Claude or ChatGPT into Salesforce will already be paying those providers, and Flex Credit consumption introduces another cost to account for on the Salesforce side.
Tim also questioned the scope of the model, noting that it applies not just to Salesforce-hosted MCP access, but to agentic API usage more broadly.
“Organizations will need to know what they’re looking to achieve with their agents, and estimate the cost of running them before they can make a decision on whether or not to proceed,” he said.
Final Thoughts
The security case for giving agents their own identities and permissions might be relatively clear, but the harder question at the moment will be the cost factor.
Because the model applies to agentic API usage as well as MCP, customers using external AI tools could face charges from both their AI provider and Salesforce. Until the Flex Credit multiplier is published, though, it is difficult to know how significant that impact will be.
