Ahead of Dreamforce 2026, Salesforce introduced something new to the Agentforce lineup: names. On September 11, the company announced seven "job-ready" agents, each built for a specific role rather than a blank canvas you configure yourself. By the time the keynote opened in San Francisco, the team had a full introduction on stage.
Meet the team: Casey, Paige, Carter, Marshall, Piper, Fin, and Hunter.
What Each Agent Actually Does
Casey handles customer service across voice, SMS, WhatsApp, and web chat, with pre-built support for FAQs, returns, account management, and escalation to a human when needed. Picture a customer messaging on WhatsApp to return a damaged item: Casey checks the order, confirms it qualifies, generates the return label, and only pulls in a human if the customer pushes back on the resolution.
Paige is built for IT and HR. It resolves employee requests inside Slack, internal portals, and the tools staff already use. Say an employee asks in Slack why their expense report is stuck in approval: Paige checks the workflow, tells them exactly who it's waiting on, and can nudge that approver directly. Autism Queensland reportedly uses Paige to handle 70% of its administrative requests.
Carter works commerce. It helps shoppers discover products, compare options, and complete checkout inside the chat itself, so a customer comparing two similar products can get specs, check stock, and pay without ever leaving the conversation.
Marshall runs back-office and supply chain processes, like supplier onboarding, with deterministic execution and an audit log for every action. Instead of a person working through a multi-week intake, compliance, and system-entry checklist by hand, Marshall works it end to end. In one keynote demo, a Salesforce presenter said Marshall "learned" a new onboarding process in 90 minutes inside an SAP sandbox, a task that would typically take a new employee weeks to pick up.
Piper works inbound pipeline: engaging prospective buyers and qualifying them into real opportunities without waiting on a human follow-up. A visitor fills out a "contact us" form; Piper follows up immediately, asks the qualifying questions a rep normally would, and only hands the lead over once it's worth a rep's time.
Fin handles more complex customer experience conversations, the kind that span more than one issue. If a customer's question touches both a billing problem and a technical one, Fin can work through both in the same conversation instead of routing them between departments. It joined the lineup through Salesforce's completed acquisition of Intercom, announced September 10, and is now part of Salesforce AI Labs. Anthropic, an early customer, reports Fin resolves 79% of its support conversations.
Hunter is the outbound sales agent, working a pipeline from research through outreach and staying on a deal for weeks or months. It researches a target account, drafts personalized outreach, follows up as the deal moves, and checks in with the seller before anything that needs a judgment call. One customer, Perk, reports that 60% of its sales pipeline is now built by Hunter.
Six of the seven, everyone except Hunter, are generally available now. Hunter is in pilot, with general availability planned for November 2026.
Why Named, Role-Based Agents Matter
Most agent platforms, Salesforce's included, have led with a builder: a blank canvas and a set of tools to assemble an agent yourself. That approach works, but it puts the burden of design on the customer. This release inverts it. Instead of starting from nothing, a company can start from a role already scoped to a job, a set of actions, and a stated availability date.
That's a meaningful shift for adoption. It's the same idea Alforce ties into: getting real value out of an AI investment increasingly means matching a defined business function to an agent built for it, not evaluating a general-purpose platform in the abstract.
What This Means if You're Evaluating Agentforce
If your team is weighing Agentforce, the named lineup gives you a more concrete starting point than the platform alone. Casey, Paige, Carter, Marshall, and Piper are all live today. Fin depends on the specific support workflow it plugs into. Hunter is worth watching, but treat any near-term rollout as a phased pilot rather than a production commitment until its November GA.
Casey's job, resolving service issues against real account and order data instead of a canned script, isn't new to us. We built a version of it for Migros inside the Migros Super App: Service Cloud and Agentforce handling roughly 500,000 chats and 300,000 support tickets a month.
