Salesforce launches pay-per-resolution agent

Salesforce has launched the Agentforce Help Agent, which it says can be deployed "in minutes" and is charged based only on the resolutions delivered.

Agentforce Help Agent is the next iteration of Agentforce and following a two-year development period offers distinct USPs that tackle the challenges practitioners currently face around the practical use of AI agents: 

Pay-per-resolution pricing: Salesforce says users will only pay when the Help Agent autonomously resolves an issue from start to finish. If a customer gives negative feedback or asks for human escalation, there is no charge, and the agent passes complete customer context to the service team. Both Data 360 and Agentforce are unmetered during the agent interaction. 

Out-of-the-box knowledge and testing: Because many companies still struggle with their data and knowledge, Help Agent is grounded on the existing Salesforce Knowledge and can be connected to knowledge bases, workflows, and service channels in minutes. It also allows users to drag and drop additional files or enter a web URL for crawling. Built on the Agentforce Platform and Data 360, the agent is securely grounded in the customer's full context, allowing for immediate action, from order management to appointment scheduling and an agent preview pane enables agent testing to help ensure confidence in the responses before moving to the next step.

Prepackaged actions for rapid deployment: Eliminating the need for complex orchestration and system integration, Help Agent works across channels including voice, web, portal, and messaging from a single screen. It also comes with prepackaged actions: it can answer customer questions and manage cases, and additional actions such as order management, appointment scheduling, and account management can be added using the existing setup options in Agentforce Builder or the coding agent of the customer's choosing.

The launch comes as a Salesforce survey found the adoption of AI agents in customer service has grown from 39 percent in 2025 to 66 percent in 2026.

Kishan Chetan, EVP and GM of Agentforce Service at Salesforce said: "The promise of AI agents in customer service isn't just about answering questions faster; it's about resolving issues completely, across any channel, the first time and every time. 
"Built on our deeply unified, secure-by-design platform, it delivers a personalized, proactive omni-channel experience that is even easier to set up. And with our pay-per-resolution pricing, our success is directly tied to our customers' success," he added.  

Don't miss any news, updates or insider tips from CX Network by getting them delivered to your inbox. Sign up to our newsletter and join our community of experts. 

Learn More

What is pay-per-resolution pricing?

CX is learning the hard way that AI costs can quickly add up. Vendors offer a number of different cost models, from per-minute billing with added fees for each resolution, to mandatory annual minimum commitments tied to projected usage. 
Outcome-based pricing, where users only pay when AI successfully resolves an issue, saves a cancellation, or drives an upsell, is increasingly popular and, as outlined, this is the model offered for Agentforce.

"Salesforce is right to move away from seat-based pricing," says Bill Staikos, founder and managing parter of Be Customer Led. "The market wants to pay for work completed, not software accessed. But the harder question is who defines 'resolved', especially when a customer gives up, reopens the issue, escalates later, or leaves unhappy," he adds. 

Staikos says it remains to be seen if Agentforce can handle "messy, emotional, multi-step customer issues where context is incomplete and policy exceptions will break the AI, resting in a handoff". He says: "That's where most enterprise service costs live, and again resolution gets complicated."

He continues: "At the end of the day, AI makes access-based pricing much harder to defend. Salesforce is just moving with the market here anyway; it's nothing new. Just look at the Voice AI space that's eating their lunch, for example, Decagon, Sierra, etc. If it doesn't work, enterprise customers will learn very quickly that 'resolved' and 'valuable' aren't the same thing."

The $3.6bn Fin deal

The new Agentforce Help Agent news comes only days after Salesforce's acquisition of Fin. Formerly known as Intercom, Fin is a customer agent platform for SMEs, that provides autonomous, end-to-end AI service agents. Fin is used by more than 30,000 SMEs globally.

It's the latest strategic acquisition in Salesforce's growth strategy, following its acquisitions of Contentful, M3ter, Informatica, Slack, Spindle AI and Momentum, and others. 

This deal gives Salesforce new capabilities in AI agent deployment for enterprises. Fin's AI Agent is powered by its proprietary AI model, Apex. It's purpose-built for customer support and has demonstrated industry-leading resolution rates – claiming an average 76 percent end-to-end – that outperform other commercially available frontier models.

Fin's USPs include faster time-to-value options for scenarios where a business needs to launch quickly, integrate with existing systems, and deliver measurable outcomes.
Salesforce said the technology "will help organizations improve autonomous resolution, reduce cost-to-serve, and accelerate AI adoption across their service organizations".

The acquisition will also bring a long-standing technical AI team and an established global customer base of more than 30,000 companies to Salesforce.

CEO and co-founder Eoghan McCabe said: "Our technology has defined this category and set the new standards for what great customer service looks like today. By joining forces with Salesforce, we can deploy it far and wide at a rate far faster than we could have ever achieved on our own."

The US$3.6 billion deal was confirmed on June 15. The transaction is expected to close in the fourth quarter of Salesforce's fiscal year 2027, subject to satisfaction of customary closing conditions, including the receipt of required regulatory clearances.

Quick links

• Refinanced Medallia to double down on generative AI and automation
• Qualtrics closes Press Ganey Forsta deal for $6.75bn
• Bank of England speculation sparks backlash against dynamic pricing