Salesforce rolls out Agentforce Help Agent with guided setup and pay-per-resolution pricing

Salesforce on June 25 introduced Agentforce Help Agent, an autonomous customer-service agent aimed at handling support issues from start to finish, using guided setup intended to get enterprises live in minutes. The company also positioned a new pricing approach around outcomes, saying customers pay only when the agent resolves an inquiry autonomously.

Agentforce Help Agent is built on Salesforce’s Agentforce 360 Platform, which Salesforce describes as the foundation for building, deploying and governing AI agents. In this case, Salesforce says the Help Agent is focused on a single job: resolving customer issues completely, on day one, across channels where customers already contact support.

Salesforce said the product is designed to solve three common failure points for service agents. First, it grounds the agent automatically on Salesforce Knowledge, so the system is connected to the business’s own documentation. Second, Salesforce said it addresses messy data, which it called a leading reason agents fail. Third, it provides a library of available actions that can run on real workflows, including managing cases, scheduling appointments and updating orders.

A key part of the announcement is multi-channel coverage. Salesforce said Help Agent can operate across voice, web, portals and messaging, with customers turning it on from a single screen. In addition to the product’s own design, Salesforce pointed to internal usage, where it said the agent has handled 4.3 million inquiries and resolved 70 percent of them.

Salesforce also tied the economics to results. Under pay-per-resolution pricing, Salesforce said customers are charged only when the agent resolves an issue from beginning to end. If the customer asks for a human or becomes dissatisfied, Salesforce said there is no charge, framing the cost structure as linked to successful outcomes rather than activity volume.

The announcement came alongside continued expansion of Salesforce’s agent strategy. Salesforce said it recently signed a definitive agreement to acquire Fin, a customer agent platform built for small and mid-sized businesses that provides autonomous, end-to-end AI service agents. Salesforce said Fin is trusted by more than 30,000 companies globally to resolve customer queries, and the company expects the transaction to close in the fourth quarter of Salesforce’s fiscal year 2027, subject to customary conditions and regulatory clearances.

Salesforce’s own messaging suggests it is trying to reduce the friction enterprises face when rolling out AI agents, especially when they must integrate with existing systems and customer touchpoints. For established Salesforce customers, grounding on Salesforce Knowledge and reusing real workflows for actions are meant to cut the time between “model demo” and “support tool in production.” For businesses that lack dedicated agent-building teams, guided setup and multi-channel deployment are positioned as a way to shorten time to value.

Service leaders cited in the release reflect that theme. PenFed Credit Union’s Nicole LaCamp said the approach is freeing teams to focus on anticipating member needs, and she linked the direction to an outcomes-based model where both sides win. Fisher & Paykel’s Rudi Khoury said the agent recognizes the customer and context to redefine the support journey, and he claimed agentic AI has doubled self-service resolution rates. PowerSchool’s Mark Miller said the company plans to deploy frictionless, personalized service across channels to help customers reach instant resolutions.

Several details remain outside the public announcement. Salesforce did not provide pricing rates, eligibility rules for pay-per-resolution billing, or specific performance constraints across different industries, languages, or support categories. It also noted that the article may reference features still in development and unreleased, meaning some capabilities implied by the broader Agentforce roadmap may not yet be available under general availability for every customer. For buyers, the next step will be confirming what is fully released, how quickly guided setup works in their environment, and which workflow actions are supported out of the box versus requiring additional configuration.

What to watch next is whether Agentforce Help Agent’s reported results translate consistently for external deployments, especially across voice and messaging where integrations and escalation handling can be more complex. Investors and customers will also be watching how the Fin acquisition progresses and how quickly Salesforce can broaden deployment options for small and mid-sized businesses. Finally, the pay-per-resolution approach may become a bellwether for how agent vendors define “success” and align incentives as enterprises scale autonomous support.