Salesforce shifts to consumption pricing as AI 'Flex Credits' drive half of bookings
The CRM giant is moving away from seat-based licenses toward a usage-driven model to monetize its AI agent ecosystem.
Salesforce is fundamentally altering its revenue engine as it integrates generative AI into the enterprise workflow. The company is moving away from traditional seat-based licensing in favor of a consumption-driven model that allows it to capture value from AI usage spikes.
In its second-quarter results, Salesforce reported revenue of $11.3 billion, marking an 11% increase year-over-year. The standout metric was the surge in consumption-based bookings. According to Miguel Milano, President and COO at Salesforce, 50% of bookings came from customers "refilling the tank" by purchasing additional Flex Credits.
To support this shift, Salesforce is diversifying its monetization strategies. The company now offers a range of pricing options for AI services, including basic usage, consumption-based models, and outcome-based payments tied to specific transactions or business results. Marc Benioff, Co-CEO and Founder, noted that this diversity in pricing has expanded the company's ability to secure very large transactions with its customer base.
The shift to 'headless' software
This strategic pivot comes as Salesforce navigates a broader industry shift where AI models interact directly with data, potentially bypassing traditional user interfaces. To counter the threat of AI replacing standard business software, Salesforce is embracing "headless" monetization and agentic workflows.
A primary example of this strategy is the launch of "Claudeforce," a partnership with Anthropic. The integration brings Claude's AI capabilities into the Salesforce ecosystem via a plugin that features 37 prebuilt "sales skills," allowing AI agents to perform complex tasks without requiring a human to navigate a dashboard.
Market implications and risks
The transition to Flex Credits represents a significant departure from the predictable recurring revenue of the SaaS era. While the high volume of "refill" bookings suggests that AI consumption is scaling faster than anticipated, the model introduces new variables for the enterprise buyer.
Industry analysts have raised concerns regarding the stability of this approach. Gartner has warned that Flex Credit models may involve upfront commitments and rates that vendors can change unilaterally, potentially creating budget unpredictability for corporate IT departments. Despite these risks, current data suggests that enterprises are willing to trade predictability for the scalability of AI agents.
What to watch
As Salesforce continues to roll out Agentforce and its various partnerships, the industry will be watching whether the consumption model remains sustainable as AI efficiency improves. The key question remains whether outcome-based pricing—where Salesforce only profits when a specific business goal is met—will eventually supersede the current credit-based system. For now, the "refill" trend indicates a strong initial appetite for agentic AI in the enterprise.