Agentic AI

Gartner questions whether Salesforce AI buffet will stay all-you-can-eat forever

Analysts say today's capped deals may become tomorrow's cost shock

UPDATED Gartner is warning Salesforce users that a capped enterprise agreement for its AI and data platforms may not be available when they come to renew, potentially meaning customers could struggle to predict costs and understand value.

Salesforce announced the Agentic Enterprise License Agreement (AELA) last October. It creates an all-you-can-eat approach to buying and consuming AI agents at a capped cost, and accompanies existing consumption-based and flexible credit models.

Speaking to The Register, Hannah Decker, Gartner director analyst for IT sourcing and procurement, said it is critical that users understand the exit terms of AELAs before signing up for them.

"Gartner believes that these are going to be converted into defined quantity contracts at the end of the agreement," she said.

Salesforce told The Register it would not move away from capped agreements and that renewals would remain flexible, and because AI compute costs may actually shift over time.

Salesforce customers with unlimited AI access should clarify how long that arrangement will last. "If they're moving to a defined quantity contract, there needs to be limits on price increases at renewal," Decker said. "Making sure there are caps that protect you when the agreement ends is critical."

Late last year, Salesforce's chief revenue officer said that he is relaxed about the CRM giant losing money on AELAs in the short term because the corporation will have many more years to "monetize" customers. Miguel Milano told investors: "The customer deploys so much that all of a sudden, that deal is not profitable for me... And then I have another 20 years to monetize that customer."

Decker said that even before AELAs come to an end, users should try to understand usage and manage it where appropriate, so they know what they need when the terms end.

"There does need to be some internal scrutiny around that, and ensuring that there's also value aligned with the usage and that customers are ready for that. Value is not necessarily being measured effectively at this point, so this is certainly an area that organizations need to pay attention to."

There is also the challenge of predicting costs in defined quantity contracts, as users struggle to know how much they need. Because agentic AI is a new capability, there's a lack of historic data to back up forecasts, Decker said.

"Buyers are really struggling with that. There's additional risk to buyers in that, in most cases, vendors retain the right to change the multiplier, or the rate of consumption of those credits at any point during the term of the agreement, which basically means that consumption could increase even though usage hasn't changed, which means higher costs for unchanged usage. So there's also that additional element of risk."

In a research note, Gartner warned that proposed price increases from SaaS vendors are above the rate of inflation, as vendors look to drive profitability higher or to cover the cost of GenAI development. It warned that Salesforce price rises of 6 to 15 percent would feed directly into renewal uplifts, with GenAI and agentic AI SKUs "leading to steep cost increases."

Decker said: "There is a lot of confusion around how these models work, and that's not just isolated to Salesforce. With consumption models, because of the inconsistency from vendor to vendor, many buyers are really struggling to understand how they can control costs."

Last year, Salesforce CEO Marc Benioff told investors he saw a "very high margin opportunity" in users adopting its AI agent platforms. Customers just need to make sure this does not come at the cost of their budgets. ®

Updated at 12.55 UTC on January 29, 2025, to add:

Salesforce contacted us following publication of this article to contest the statements made by Gartner, as now included in paragraph five.

In a statement sent to The Register, Bill Patterson, Salesforce EVP, Corporate Strategy, said:

"The claim that we are moving away from capped agreements is inaccurate. Renewals remain flexible, and because AI compute costs may actually shift over time, we focus on tailoring terms so each customer can get the maximum value from their usage. And our Digital Wallet tool gives customers real-time visibility into their AI and data spend, enabling them to forecast their AI and data investments accurately.

"Our consumption based pricing model for AI and data is specifically tied to business outcomes, and allows customers to start small, experiment, and scale across the enterprise over time. This flexible model ensures that an organization can shift investments according to priorities, and has been well received by customers."