Salesforce expanded Agentforce on September 14, 2026 with named, job-titled agents, Casey in support, Paige in IT and HR, Carter in shopping, Marshall in supply chain, Piper in inbound lead qualification, Fin in customer service, and Hunter in outbound sales, each customizable for skills, actions, data models, permissions, and operating rules (Newsbytes.ph, September 2026), but the pricing underneath the names has not changed shape. Salesforce's own pricing page lists an Agentforce user license at $5 a month, add-on licenses at $125 to $550 a user a month depending on the cloud, and consumption pricing on top: a $2 charge for each customer-facing conversation, or $500 for 100,000 flex credits (Salesforce, 2026). Giving an agent a name and a job title is a branding decision. The bill is still built around how many people use it and how much it runs.
Key Takeaways- Salesforce's named agents, Casey, Paige, Carter, Marshall, Piper, Fin, and Hunter, launched September 14, 2026; all but Hunter (in pilot, general availability November 2026) are already generally available (Newsbytes.ph, September 2026).- Agentforce pricing combines a per-user license, from $5 a month for the base license up to $550 a user a month for the bundled edition, with consumption pricing on top: per-conversation charges or flex credits (Salesforce, 2026).- Insight: A named agent with a job title reads like a hire. The invoice underneath reads like a metered utility bill: a license line for every person who can use it, plus a usage line for every time it runs.- The biggest CRM vendor on earth adopting "AI team" language validates the category we describe in how to get an AI team for your small business. It also raises the bar: the framing is now common, so what a vendor actually charges for, and what it actually delivers, matters more than the name on the badge.- We do not itemize a client's login count or activity volume. One retainer, one role, scoped in a written proposal.
Key Takeaways
- Salesforce's named agents, Casey, Paige, Carter, Marshall, Piper, Fin, and Hunter, launched September 14, 2026; all but Hunter (in pilot, general availability November 2026) are already generally available (Newsbytes.ph, September 2026).
- Agentforce pricing combines a per-user license, from $5 a month for the base license up to $550 a user a month for the bundled edition, with consumption pricing on top: per-conversation charges or flex credits (Salesforce, 2026).
- Insight: A named agent with a job title reads like a hire. The invoice underneath reads like a metered utility bill: a license line for every person who can use it, plus a usage line for every time it runs.
- The biggest CRM vendor on earth adopting "AI team" language validates the category we describe in how to get an AI team for your small business. It also raises the bar: the framing is now common, so what a vendor actually charges for, and what it actually delivers, matters more than the name on the badge.
- We do not itemize a client's login count or activity volume. One retainer, one role, scoped in a written proposal.
Salesforce gave Agentforce seven named, job-specific agents, each with a defined function: Casey handles customer support across voice, SMS, WhatsApp, and web chat; Paige handles IT and HR employee support; Carter is a shopping assistant for product discovery and in-chat checkout; Marshall automates supply-chain processes; Piper qualifies inbound B2B leads; Fin runs customer-service workflows; and Hunter, still in pilot with general availability set for November 2026, handles outbound account research and sales outreach (Newsbytes.ph, September 2026). Every one of them is customizable on names, skills, actions, data models, permissions, and operating rules, so a business is not stuck with the out-of-the-box version.
That is a real product expansion, not just a marketing refresh, and it is worth taking seriously as a signal: the largest CRM vendor in the world has decided "a named AI employee doing one job" is the framing that sells.
No, and the gap between the framing and the invoice is the part worth understanding before adopting any of it. Salesforce's own pricing lists a base Agentforce user license at $5 a month per user, add-on licenses running $125 to $150 a user a month for specific clouds, and full Agentforce 1 editions starting at $550 a user a month, layered with consumption pricing on top of the license: a $2 charge for each customer-facing conversation, or roughly $500 for 100,000 flex credits, with each agent action consuming a slice of that pool (Salesforce, 2026). An "employee" with a name still requires a license for every person who touches it and a running meter for every action it takes.
That is a coherent business model for Salesforce. It is also the exact structure a buyer needs to see clearly before assuming a named agent behaves like a flat-cost hire: the name describes the job, not the bill.
Because "AI team" framing is now mainstream, coming from the biggest CRM vendor on earth, and mainstream framing raises the bar for what actually backs it up. When a category everyone recognizes gets a familiar name attached, "AI employee," "AI team," the buyer's job gets harder, not easier: the words now signal less about what is actually being sold, because every vendor uses them. What still differs vendor to vendor is what happens after the announcement, how the thing is actually billed and what it actually does when the standard configuration does not fit.
That is the same distinction we cover in our flat-retainer pricing explainer: the label on the product is not the same claim as the structure of the bill, and a small business evaluating either has to check both.
Whether the total cost scales with headcount and activity, or is fixed to the scope of the job. A license-plus-consumption structure like Agentforce's means the bill grows as more people in the business touch the tool and as the agent does more work, which can be fine for a business with predictable, bounded usage and hard to plan around for one that is not. An embedded AI team role built around a specific function, back office, operations, lead qualification, account management, or outbound, runs on one retainer scoped to that function, not metered by how many people log in or how many actions the role takes in a given month.
Neither structure is wrong on its face. The mistake is assuming a friendly name and job title means the pricing behaves like a flat-cost hire when it has not been checked against the actual pricing page.
This distinction holds across verticals too: whether the role is chasing invoices in accounting or qualifying leads in real estate, see the industries pages for how a scoped, single-retainer role compares to a licensed, metered platform in practice.
Functionally adjacent, not directly comparable. Agentforce is a licensed, consumption-priced platform a business configures and runs inside Salesforce's ecosystem. We build and run AI team roles scoped to a specific function, on one retainer, inside or connected to whatever CRM a business already uses, Salesforce included.
The license price per user is the same regardless of company size; what differs is how many users and how much consumption a business needs, which determines the total. A five-person team pays fewer license lines than a five-hundred-person one, but the per-unit structure, license plus consumption, is identical.
Salesforce's public pricing page does not break out a separate price for each named agent; the license and consumption-credit structure applies across the Agentforce product line, with the specific mix depending on which clouds and features a business enables.
Because the framing shift, AI as a named "employee," is becoming the default marketing language across the category, and a buyer evaluating any vendor using that language, including us, should know to check the billing structure underneath it rather than take the framing at face value.
Salesforce naming its agents is a real signal that "AI team" language has gone mainstream. What has not changed is the discipline a buyer needs: check what actually gets billed, per user, per action, or per scoped function, before deciding what you are buying. Send a written intake if what you want is the third option, and the proposal that comes back is one number for the function, not a license count and a credit meter.
Internal links to add from older posts within a week: anchor "AI pricing built on usage and license counts" pointing to why-we-dont-charge-per-token-or-seat; anchor "AI team roles" pointing to how-to-get-ai-team-for-small-business; anchor "productized agent platforms" pointing to no-code-agent-builders-vs-custom-ai-engineer.
Maxpertise is an AI-native engineering company. We embed native AI engineers inside your team, live in about 10 days. Please enable JavaScript to view the site, or email contact@maxpertise.net.