The short version
- Salesforce announced seven named, prebuilt 'job-ready' agents on September 11, 2026, just ahead of Dreamforce: Casey, Paige, Carter, Marshall, Piper, Fin, and Hunter. Six were generally available at announcement; Hunter is in pilot with GA planned for November.
- Each ships with the skills, actions, and data models for a specific job, so a team starts from a working agent rather than a blank canvas, and customizes it with Agent Script. Agentforce Coworker is a separate, general-purpose teammate, not one of the seven.
- Fin is the Intercom Fin: Intercom rebranded around its agent, Salesforce signed a roughly $3.6 billion acquisition in June, and the deal closed on September 10, one day before this roster was announced.
- Salesforce has published no official mapping of which agents come with which editions, so treat per-agent licensing as a scoping question, not a datasheet fact.
What "Job-Ready" Actually Means
On September 11, 2026, days ahead of Dreamforce, Salesforce announced a portfolio of seven prebuilt Agentforce agents with human names and specific jobs. The pitch: instead of assembling topics, actions, and instructions from a blank canvas, you deploy an agent that already ships with the skills, actions, and data models its job requires, wired into Customer 360, and you customize it with Agent Script, the open-source language that mixes AI reasoning with deterministic rules.
Two boundary lines keep the family straight. First, Agentforce Coworker is not one of the seven: it is the separate, general-purpose AI teammate inside Lightning that can orchestrate other agents. Second, these are packaged employees, not a new platform; everything under them, the runtime, orchestration, and governance, is the same Agentforce your custom builds use. If you are new to that platform, start with our getting started guide.
Six of the seven were generally available at announcement. Hunter, the most ambitious, is in pilot with general availability planned for November 2026. Here is the roster, one by one.
Casey: The Help Agent
Casey resolves customer service issues across voice, SMS, WhatsApp, and web chat, with prebuilt patterns for FAQs, returns, account management, and escalation to a human. Worth knowing: Salesforce Ben reports Casey is the renamed Agentforce Help Agent that launched mid-2026, and the lineage matters because the Help Agent introduced pay-per-resolution pricing at $2 per resolved conversation, with no charge when it escalates. If service deflection is your use case, Casey is the shortest path in the roster.
Paige: The IT and HR Agent
Paige handles employee requests, password resets, access grants, HR questions, inside Slack, portals, and the tools your people already use. Salesforce cites Autism Queensland routing 70% of administrative requests through Paige (a vendor claim, like every proof point in this post). For teams drowning in internal tickets, Paige competes directly with standalone ITSM bots, with the advantage of living where your employee data already is.
Carter: The Shopper Agent
Carter works the commerce funnel: product discovery, comparisons, question-answering, and conversion through in-chat checkout. Salesforce cites retailer Hibbett covering 90% of core shopper journeys with a six-week deployment. If you run Commerce Cloud, Carter is the agent that turns the chat widget from a cost center into a register.
Marshall: The Supply Chain Agent
Marshall orchestrates back-office supply chain processes end to end, and its distinguishing feature is temperament: deterministic execution with audit records, because procurement and logistics are places where an improvising agent is a liability. It is the least flashy of the seven and the clearest signal that Salesforce intends agents for regulated, auditable work, not just chat.
Piper: The Inbound Pipeline Agent
Piper engages inbound leads on your website and inbox, qualifies them, and converts them into pipeline before a human touches the deal. Salesforce cites Asana seeing four times the conversation volume with 45-day average deployments. Piper is the natural first agent for B2B marketing teams whose forms and chat currently leak qualified interest.
Fin: The Customer Agent, and the $3.6 Billion Story
Fin resolves complex customer experience workflows across every channel, powered by the Operator and Fin Apex models. The name is not a coincidence: this is Intercom's Fin. Intercom rebranded the whole company around its agent, Salesforce signed a definitive agreement in mid-June 2026 at roughly $3.6 billion, and the acquisition closed on September 10, one day before this roster was announced. Fin continues serving its 30,000+ existing customers from inside Salesforce AI Labs, and Salesforce cites a 76% average resolution rate.
One tension to watch as the pricing stories merge: before the acquisition, Fin charged roughly $0.99 per resolution while Salesforce's Help Agent charges $2, per secondary reporting. How that reconciles is not yet published, which is exactly the kind of question to put in a scoping call rather than assume.
Hunter: The Outbound Agent on a New Runtime
Hunter works outbound sales pipeline from account research through outreach, drafts, presentations, and proposals, and it persists: Salesforce describes it staying on a deal for weeks or months alongside a seller. It is the first agent built on the new long-horizon runtime, which adds memory, durable execution, and dynamic steering so an agent can pursue an objective rather than complete a task. Salesforce cites early customer Perk attributing 60% of its sales pipeline to Hunter. It is in pilot now, with GA planned for November 2026.
Pricing, Honestly
Here is what the announcement does not contain: pricing, or any official mapping of which agents come with which editions. Do not let anyone assert one to you; as of late September it is not published.
The context that does exist: the Help Agent lineage charges $2 per resolved conversation. Agentforce actions otherwise draw on Flex Credits at $500 per 100,000 credits, roughly $0.10 per action and more for voice. And the new Core, Advanced, and Max editions announced September 3 ($195, $395, and $550 per user per month) bundle Slack, Tableau Next, Agentforce Coworker, and escalating Flex Credit allowances, but their announcement does not name the seven agents either. Model your economics with the Agentforce ROI calculator, then verify the license path on a real quote.
Which One to Pilot First
The honest heuristic: pick the agent whose job your team already measures, because you will know within a month whether it earns its keep.
- High-volume support queue: Casey, and measure deflection against the $2-per-resolution economics.
- Internal IT and HR tickets eating your ops team: Paige, especially if you live in Slack.
- Commerce site with an underperforming chat widget: Carter.
- B2B site leaking inbound interest: Piper, the fastest of the seven to show revenue impact.
- Outbound-heavy sales motion: wait for Hunter in November rather than forcing an inbound agent into an outbound job.
- Existing Intercom shop or CX at scale: Fin, which you may already be running.
- Supply chain and procurement: Marshall, and treat its audit trail as the feature you are actually buying.
The Honest Read
The skeptical context matters. Salesforce Ben pegged Agentforce adoption at roughly 12% of customers earlier in 2026, and analysts framed this cycle as a simplification-and-adoption story rather than a capability leap: named agents with jobs are easier to buy than a platform with potential. That is not a criticism; packaging is how enterprise software actually gets adopted. But every proof point above is a vendor claim, the agents still inherit your data quality and your permission hygiene, and a prebuilt agent mis-scoped is just a faster way to disappoint stakeholders.
Our read: the roster is real progress, Piper and Casey are the lowest-risk first moves, Hunter is the one to watch in November, and the deciding factor is not which agent you pick but whether your org is ready for any of them. That readiness work, data, consent, escalation paths, is what we do. Bring your queue volumes and your org to a free 30-minute scoping call at cal.com/cloudsheer-consulting/30min and we will tell you which of the seven fits, or whether none of them do yet. Our Agentforce practice takes it from there.