Salesforce Gave Its AI Agents First and Last Names: The Digital Employee Era | inicIA

Direct answer

On September 11, 2026, Salesforce launched Agentforce 360: seven AI agents with their own names and a defined business role —Casey (customer service), Paige (internal IT and HR support), Carter (commerce), Hunter (outbound sales), Marshall (supply chain), Piper (inbound pipeline), and Fin (customer experience). Alongside them it introduced a Control Plane to govern agents and a long-horizon runtime that lets agents pursue goals across days and weeks, not just a single conversation. The lesson for your business: the era of the generic chatbot is over. What works are agents with a clear role, real data, and defined supervision.

The 7 digital employees

Each agent is built for one function, not to "do everything":

Six are available at launch; Hunter is the exception because Salesforce wants more validation before handing an agent direct control of outbound selling. That detail says a lot: even the most aggressive giant applies the brakes where commercial risk is highest.

The names are not the point: the Control Plane is

The announcement that truly matters is not the name list. It is the AI Control Plane: a layer to register, govern, and monitor AI agents —Salesforce's and third parties' (Amazon Bedrock, Microsoft Foundry, OpenAI, Gemini)— from a single surface. It controls each agent's identity, policies, lifecycle, performance, and cost.

The real enterprise problem is no longer a shortage of agents: it is having dozens of ungoverned agents. If your business is going to use AI seriously, the first question is not "which agent do I buy" but "who supervises it, under what rules, and how do I measure its performance." Salesforce understood that and turned it into a product.

The long-horizon runtime: agents that work for weeks

The other big piece is the long-horizon runtime: agents can now pursue a goal across days and weeks instead of resolving a single interaction. It rests on three capabilities: memory (context is preserved across sessions), durable execution (multi-step plans persist over time), and dynamic steering (the agent adjusts its actions with your feedback). Hunter is the first agent running on this runtime.

For your business, this changes the conversation: an agent is no longer "the one answering the chat," it is "the one working the October pipeline until it closes, under your supervision."

What to copy in your business today

You do not need Salesforce to apply the lesson. Three principles:

  1. One agent, one role. Define the job before switching anything on: what it handles, what it never touches, and when it asks a human for help.
  2. One human owner. Every agent needs an accountable owner who reviews its performance the way they would review an employee.
  3. Outcome metrics. Salesforce reports 7 billion agentic work units delivered (3.2 billion in Q2 alone) and cases like Perk, which credits Hunter with 60% of its pipeline. Those are vendor figures, not independent audits — take them as direction, not promise. What you can copy is the habit: measure by outcome, not activity.

Related articles

Frequently asked questions

Does this replace my employees?

No: it replaces them in specific, repeatable tasks. The pattern that works is AI + human: the agent handles volume and your team handles judgment, relationships, and closing. Salesforce itself keeps Hunter in pilot precisely where commercial judgment is critical.

Can a small business use something like this?

Salesforce's platform is enterprise, but the model scales down: a WhatsApp agent that qualifies and books, a voice agent that answers calls, an agent that does follow-up. One role, one owner, clear metrics.

What is the Control Plane in simple terms?

The "boss of the agents": who each agent is, what it is allowed to do, what it costs, and how it is performing. If you will run more than one agent, you need something like this even if it starts as a manual checklist.

How much does Agentforce 360 cost?

Salesforce published no simple pricing in the announcement; its model is metered in agentic work units on its platform. When evaluating vendors, always ask for cost per outcome, not per license.

Take the next step

If you want agents with defined roles for your business —handling, qualifying, and following up under clear supervision—, talk to inicIA on WhatsApp: https://wa.me/17866120930

By inicIA