Salesforce introduces outcome-based pricing for Agentforce AI

Salesforce introduces outcome-based pricing for Agentforce AI

The CRM giant will charge customers $2 per successful AI resolution, only billing when the bot actually solves the problem

Share

Add us on Google
by Editorial Team

Salesforce just flipped the script on how enterprise AI gets sold. Instead of charging customers for access to its Agentforce platform regardless of whether it works, the company is now offering a pricing model where customers only pay when the AI actually delivers results.

The new outcome-based pricing applies to the Agentforce Help Agent and costs $2 per successful autonomous resolution. If the AI doesn’t solve the customer’s problem, Salesforce doesn’t get paid.

What counts as a successful resolution

A billable resolution requires at least two message turns between the AI agent and the customer. The interaction also needs to receive non-negative feedback from the user. And critically, if the conversation gets escalated to a human agent, no charge applies.

Advertisement

For customers who prefer traditional consumption metrics, the same resolution can alternatively be purchased at 400 Flex Credits, with flex credits priced at $0.10 per action. Salesforce also maintains a per-user licensing option, giving enterprises three distinct ways to pay for the same product.

The numbers behind the confidence

Salesforce has reported that Agentforce autonomously resolved roughly 70% of 4.3 million customer inquiries. That’s about 3 million conversations handled without a human stepping in.

The financial results back this up. Agentforce has reached an annual recurring revenue of $1.5 billion, representing more than 240% growth year-on-year.

CEO Marc Benioff has framed outcome-based pricing as a strategic lever for closing larger deals. The logic is straightforward: when vendor fees are correlated with measurable business results like revenue increases or cost reductions, procurement teams face less internal resistance to signing bigger contracts.

A broader industry shift

Competitors across the enterprise AI landscape are experimenting with similar pricing structures designed to make ROI more visible and immediate for buyers.

For Salesforce specifically, the new pricing model sits alongside its existing flex credit and per-user options rather than replacing them. A Fortune 500 company with predictable AI workloads might prefer per-user licensing, while a mid-market firm experimenting with AI customer service might gravitate toward paying per resolution.

The availability timeline puts this in production starting July 2026.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
AI
Salesforce introduces outcome-based pricing for Agentforce AI
Salesforce introduces outcome-based pricing for Agentforce AI

The CRM giant will charge customers $2 per successful AI resolution, only billing when the bot actually solves the problem

by Editorial Team

Share

Add us on Google

Salesforce just flipped the script on how enterprise AI gets sold. Instead of charging customers for access to its Agentforce platform regardless of whether it works, the company is now offering a pricing model where customers only pay when the AI actually delivers results.

The new outcome-based pricing applies to the Agentforce Help Agent and costs $2 per successful autonomous resolution. If the AI doesn’t solve the customer’s problem, Salesforce doesn’t get paid.

What counts as a successful resolution

A billable resolution requires at least two message turns between the AI agent and the customer. The interaction also needs to receive non-negative feedback from the user. And critically, if the conversation gets escalated to a human agent, no charge applies.

Advertisement

For customers who prefer traditional consumption metrics, the same resolution can alternatively be purchased at 400 Flex Credits, with flex credits priced at $0.10 per action. Salesforce also maintains a per-user licensing option, giving enterprises three distinct ways to pay for the same product.

The numbers behind the confidence

Salesforce has reported that Agentforce autonomously resolved roughly 70% of 4.3 million customer inquiries. That’s about 3 million conversations handled without a human stepping in.

The financial results back this up. Agentforce has reached an annual recurring revenue of $1.5 billion, representing more than 240% growth year-on-year.

CEO Marc Benioff has framed outcome-based pricing as a strategic lever for closing larger deals. The logic is straightforward: when vendor fees are correlated with measurable business results like revenue increases or cost reductions, procurement teams face less internal resistance to signing bigger contracts.

A broader industry shift

Competitors across the enterprise AI landscape are experimenting with similar pricing structures designed to make ROI more visible and immediate for buyers.

For Salesforce specifically, the new pricing model sits alongside its existing flex credit and per-user options rather than replacing them. A Fortune 500 company with predictable AI workloads might prefer per-user licensing, while a mid-market firm experimenting with AI customer service might gravitate toward paying per resolution.

The availability timeline puts this in production starting July 2026.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.