Salesforce wants to charge for AI outcomes, but first it needs to figure out how

Nothing demonstrates how deeply Fortune 500 companies depend on Salesforce quite like a global outage lasting more than seven hours. The interruption struck during the second day of last week's Dreamforce conference, where the $40 billion-a-year SaaS vendor was pitching AIforce and a growing collection of AI products. The question now is how the company will charge for them. AIforce illustrates why the per-user licensing model is becoming harder to sustain. Unveiled at Dreamforce, it uses Salesforce's Headless Toolkit to make the company's data, workflows, and business logic available through interfaces including Slack and Claude. When AI agents and APIs perform the work instead of named human users occupying seats, charging per user becomes a less natural fit. Earlier this month, Bill Patterson, executive veep and general manager of CRM applications, told an investor webinar that Salesforce was devising "a new pricing structure that really aligns to the benefits that customers realize from this new technology." Under one such arrangement, AI customer service agents would be priced according to the cases they resolve. Outcome-based pricing will not suit every use case, however. Patterson acknowledged that some agents operate across multiple disciplines, domains, and products, making it difficult to identify a single measurable result. Salesforce is therefore also developing bundles and using Flex Credits to charge customers according to consumption. The company has reason to look beyond traditional licensing. License revenue is dragging on growth, while AI agents complicate an approach built around charging for human users. "Licenses are weighing things down right now. They're a headwind," Mike Spencer, Salesforce's deputy CFO and head of finance, told the Deutsche Bank 2026 Technology Conference last month. Spencer described Salesforce's experiments as "an anxiety-filled architecture right now of different pricing structures and contract frameworks." The company is testing three ways to charge: conventional seat licenses, consumption through Flex Credits, and fees tied to outcomes. Spencer expects consumption revenue to grow as more customers put AI systems into production and purchase additional credits – what the company calls "refilling the tank" – although he said it would take another three to five years to become a material share of revenue. Alongside those charging mechanisms, Salesforce offers several contract structures. Traditional annual commitments sit alongside Agentic Enterprise License Agreements (AELAs), which provide broad access for a fixed price, and the newer Salesforce Commit. Salesforce Commit resembles the model used by hyperscalers. A customer might commit to spending $10 million over three years, for example, then divide that allowance among seats, Flex Credits, and other forms of consumption as its requirements develop. Outcomes are harder to price because both sides must agree on an objective measure of success. "The key with outcome-based pricing is you've got to be very objective about the measurements that are driving the outcomes, and that always tends to be the challenge," Spencer said. For now, customers must navigate the contracts already on offer. In January, Gartner warned Salesforce users that all-you-can-eat AELAs might not be available at renewal, making future costs harder to predict. The analyst firm recommended negotiating limits on price increases if Salesforce moved customers to contracts with defined usage allowances. Salesforce denied that it was moving away from capped agreements. It said renewals would remain flexible and tailored to help customers get the most value from their usage. Preet Takkar, PwC's global and US Salesforce leader, expects outcome-based pricing to become far more prevalent by 2030, although "it will take time." PwC has a global alliance with Salesforce. Takkar said PwC had already introduced outcome-based pricing for some customers, putting its own fees at risk through gain-sharing or risk-and-reward agreements "so that our clients do not have to print new money" for digital and AI transformation. Takkar argued that seat-based licensing makes less sense when AI agents, rather than only human employees, access applications and their data. Charging for every API request could discourage use, he said, whereas completed work or business outcomes could provide more meaningful units for billing. Takkar said he did not expect capped agreements to disappear immediately, but predicted that all-you-can-eat deals would gradually give way to spending commitments and, eventually, charges tied to results. Customers should press Salesforce on the economics rather than merely demanding more features and agents, Takkar said. Mature buyers increasingly want to know how the technology will affect profit and loss and what business outcomes it will produce. Some Salesforce customers may be questioning the business outcome of this week's outage. Although service has been restored, the interruption demonstrated how dependent organizations already are on the platform – a reliance likely to deepen as they entrust more work to its AI agents. ®