Salesforce's new Agentforce portfolio covers seven roles spanning customer service and supply chains. Job-specific skills and data models arrive ready for companies to adapt to their own processes. Most options are generally available. The sales agent Hunter remains in pilot, with Salesforce planning a November 2026 general release for the option that maintains plans across sessions.
Customer-management software company Salesforce has added seven purpose-built AI agents to its Agentforce platform. Each arrives with the skills, actions and data models for a defined job, which companies can adapt to their processes. The agents connect to Customer 360, Salesforce's customer-data platform. Businesses can therefore begin with an existing role-specific agent instead of building each one from scratch.[1], [2]
Customer-service agent Casey operates across voice, text messages, WhatsApp and web chat. Paige handles employee requests in Slack and internal portals. Shopping agent Carter answers questions and supports checkout inside a conversation. Marshall runs supply-chain back-office processes while logging its actions. Piper handles inbound prospects, while Fin addresses customer-experience workflows. These roles cover separate parts of a company's customer and employee operations.[1]
Most options are generally available, but outbound-sales agent Hunter remains in pilot. Salesforce plans general availability in November 2026. Its runtime maintains a plan across sessions and allows adjustments as new information arrives. Agent Script, the company's open-source language for defining agent behaviour, also lets customers bind some decisions to fixed rules alongside model reasoning.[1]
