Salesforce's 7 Job-Ready AI Agents: A Mid-Market Hiring Guide

Last Thursday, four days before Dreamforce opened its doors, Salesforce announced a batch of new hires. None of them are human. The September 11 announcement introduced seven named, "job-ready" AI agents — Casey, Paige, Carter, Hunter, Marshall, Piper, and Fin — each built to take on a specific role, from answering support calls to building sales pipeline. It's the clearest signal yet of where Salesforce is taking Agentforce: away from "here's a toolkit, go build an agent" and toward "here's an employee, go onboard it."

For mid-market B2B teams, that shift matters. The biggest barrier to AI agents so far hasn't been ambition — it's been the build. Pre-configured agents with names, job descriptions, and reference customers change the buying decision from a development project into something closer to a hiring decision. So let's treat it like one. Here's who's on the candidate slate, which roles are worth interviewing for a mid-market B2B org, and the questions to ask before you sign an offer letter.

What Salesforce Actually Announced

The September 11 release introduces seven agents, each aimed at a defined job function. Per Salesforce's announcement, the lineup looks like this:

• Casey — Help Agent. Customer service across voice, SMS, WhatsApp, and web chat. Generally available now.
• Piper — Inbound Pipeline Generation Agent. Engages and qualifies inbound leads. Generally available now.
• Hunter — Outbound Sales Agent. Pipeline building and seller collaboration. In pilot now, with general availability slated for November 2026.
• Fin — Customer Agent. Handles complex, multi-step customer experience workflows across channels. Generally available now.
• Paige — IT & HR Service Agent. Fields employee requests through Slack and service portals. Generally available now.
• Carter — Shopper Agent. Product discovery and in-chat checkout for commerce. Generally available now.
• Marshall — Supply Chain Agent. Orchestrates back-office and supply chain processes. Generally available now.

Under the hood, the announcement leans on four platform capabilities: a long-horizon runtime that lets agents work on tasks spanning days or weeks with memory and durable execution, multi-agent orchestration so agents can hand work to each other, AI Skills for teaching agents your company's specific processes, and an Agent Optimizer for tuning performance over time. Salesforce also shared a scale number worth noting: 7 billion "Agentic Work Units" delivered to date, 3.2 billion of them in the most recent quarter alone.

Why "Job-Ready" Is the Word That Matters

The framing isn't just marketing. It targets the single biggest complaint we hear from mid-market teams about AI projects: time to value. Analyst firm Futurum, covering the announcement, points to research showing 55.1% of buyers cite faster time-to-value as the thing that gives them budget confidence. A custom agent build means months of scoping, prompt engineering, testing, and tuning before anyone sees a result. A pre-configured agent ships with the skills, data model, and guardrails for its job already in place — you configure and connect rather than construct.

Salesforce's proof points are early customers with striking numbers: Engine reporting 50% of chat inquiries fully resolved, HR platform Perk attributing 60% of its sales pipeline to Hunter, Asana citing a 45-day average Piper deployment with 4x conversation volume, and Anthropic reporting 79% of support conversations resolved autonomously by Fin. Those are vendor-published figures from reference customers, so read them the way you'd read any glowing job reference — real, but curated. The honest open question, which Futurum also flags, is whether 50–90% resolution rates replicate across the broader install base, where data hygiene and process discipline vary widely. That question gets answered in orgs like yours over the next two quarters.

The Mid-Market B2B Shortlist: Interview These Three First

Not all seven roles belong on a mid-market B2B org chart. Carter is built for consumer commerce, and Marshall assumes supply chain operations running through Salesforce. For a typical B2B company running Sales Cloud and Service Cloud, three candidates deserve a first-round interview.

1. Piper — because inbound leads are perishable

Speed-to-lead is one of the oldest problems in B2B sales, and it's brutally measurable: inbound inquiries that wait hours for a response convert at a fraction of the rate of those answered in minutes. Piper's job — engage every inbound lead immediately, qualify it, and route it — is exactly the work that falls through the cracks when your SDR team is small or nonexistent. If your web-to-lead flow currently feeds a queue someone checks twice a day, this is the highest-leverage role on the slate.

2. Casey or Fin — because support scales badly

Support volume grows with your customer base; support headcount grows with budget approvals. Casey (the Help Agent, which we covered when its pay-per-resolution pricing launched) handles the front line across chat, voice, and messaging. Fin is the heavier-duty option for multi-step workflows — think order changes, renewals gone sideways, entitlement questions that touch three systems. Most mid-market teams should start with Casey's scope and graduate to Fin's, not the reverse.

3. Hunter — worth a screening call, not an offer yet

An outbound agent that builds pipeline and coordinates with sellers is the most tantalizing role for revenue leaders — and the one still in pilot, with GA slated for November. Outbound is also where AI missteps are most visible to prospects who've never met you. Watch the pilot results, ask for B2B reference customers at Dreamforce, and plan a controlled trial for Q1 2027 rather than a quota-carrying deployment this year.

Paige, the IT and HR service agent, is a legitimate fourth candidate if your internal helpdesk runs through Slack and drowns in password resets — but for most mid-market teams the revenue-facing roles come first.

The Interview Questions: What to Check Before You "Hire"

You wouldn't hire a salesperson without an interview and a reference check. The same discipline applies here. Five questions to answer before any of these agents gets an offer:

1. Is your data ready to onboard them? A job-ready agent still learns your business from your org. If your knowledge articles are stale, your lead routing rules contradict each other, or duplicate accounts are everywhere, the agent inherits all of it — at machine speed. Fix the org first.
2. What does it cost at your volume? Casey launched with per-resolution pricing, and Salesforce is increasingly metering agent work through consumption. Before committing, model your actual monthly case and lead volumes against the pricing you're quoted, including the months where volume spikes. "Agentic Work Units" at scale is a real line item; make finance a stakeholder on day one.
3. Where does the human handoff happen? The difference between a good agent deployment and a support disaster is a cleanly designed escalation path. Define which conversations the agent must hand off, to whom, with what context — and test it before go-live, not after.
4. How will you measure the 90-day review? Set the agent's success metrics the way you'd set a new hire's ramp targets: resolution rate, lead response time, conversion on agent-touched leads, CSAT on agent-handled cases. Baseline those numbers now, before deployment, or you'll never know what changed.
5. How much orchestration lock-in are you taking on? Multi-agent orchestration is genuinely useful — and it deepens platform commitment. That's not a reason to avoid it; it's a reason to make the choice deliberately, with eyes open about switching costs later.

Watching from Dreamforce Week

Dreamforce 2026 runs today through September 17, and this agent portfolio will be all over the keynotes — expect live demos, more customer numbers, and likely additional announcements layered on top. If you're following along from your desk, the practical move is to collect claims now and validate them against your own org's reality afterward. The vendors' best week of the year is precisely the week to keep your evaluation discipline.

The Bottom Line

Salesforce packaging agents as named, job-ready hires is a real shift, and for mid-market B2B teams it lowers the barrier that has kept agentic AI in the "next year" column: the build. But job-ready doesn't mean org-ready. These agents will be exactly as good as the data, processes, and escalation paths they're onboarded into — which means the highest-ROI move this quarter may be getting your org into shape to hire them well.

That's a fast thing to find out. Run the free 2-minute org health check at 360quantumsolutions.com, or book a 5-day org audit and we'll tell you — with 16+ years in the Salesforce ecosystem behind it and a fixed fee in front of it — exactly what stands between your org and its first AI hire.