Salesforce Agentforce 2026: 7 New Job-Ready AI Agents Announced at Dreamforce

Enterprise AI has seen several developments over the past few years. From chatbots to AI assistants, humans were still needed at different steps. Salesforce’s new AI agents 2026 unveil an interesting pattern by building specific functional AI agents. 

Each AI agent is named and assigned to a specific job function, making it more like a digital employee for the users. It has a name, a job title, and performance metrics to be measured. Let us dive deep into what these agents are and how they are going to change your Salesforce environment. 

Announcement at a Glance

On September 11, 2026, Salesforce released a lineup just before Dreamforce. The release is designed around the 7 billion Agentic work units created in collaboration with Agentforce and Slack. Six of these agents are generally available today to use, but the seventh outbound sales (Hunter) agent is still in pilot and will be available around November 2026. 

Along with the AI agents, Salesforce is piloting a new governance layer of Trusted Enterprise AI Harness. This focuses on organizations that already have agents working in AI platforms but are not Salesforce native. 

7 New job-ready AI agents

Each of these seven Agentforce AI agents 2026 can be marked as Digital employees, having a job specification, name, and analytical tracking. They work with Customer 360 to get the ideal context of the data and business process. The Salesforce new AI agents 2026 are: 

Casey (Customer Service)

A help agent that works with customer service departments across voice, SMS, whatsapp, and web chat. Casey also comes with pre-built support in FAQs and returns, account management, and will hand over complex issues to humans if/when needed. 

Paige (IT & HR Support)

Paige is your IT and HR service agent that helps resolve requests across Slack and portals that employees use. It works in the self-service portal. 

Carter (Commerce Shopping)

As an e-commerce business owner, Carter is the agent you choose. It is a shopper agent that guides users to find products, compares them, and answers questions. It will help throughout the purchase journey. 

Hunter (Outbound Sales)

Hunter is the only agent still in pilot. It works a sales pipeline end-to-end. From researching accounts, preparing outreach, to collaborating with human sellers, it handles a month-long process altogether. It is the only agent built with long-horizon run time and will be available in November 2026 for general access. 

Marshall (Supply Chain)

Agentforce AI agents 2026 are not just for customer-facing operations. Marshal, the supply chain agent, works end-to-end in the back office, automating manual work with proper execution along with audit records. It is aimed directly at operations and logistics workflows. 

Piper (Inbound Pipeline Generation)

Having hefty sales teams and not getting closed deals is a big concern for enterprises. Piper is your inbound pipe gen agent that works across websites to engage, qualify, and convert inbound leads into a pipeline for B2B sales teams. 

Fin (Customer Experience)

Customer Service and experience are two different things. Where Casey supports and acts as a help agent, Fin is your customer agent. It helps handle workflows across channels by collaborating with customer operations using specialized models. It is designed for the acquisition of Fin technology and is being used at Anthropic. It is reported to resolve customer conversations with human involvement. 

Fact: You can still rename these agents to fit your branding. Agentic workflows with agents like Casey, Hunter, and Paige, Salesforce agents, will not only make your work easier but also offer faster sales and service opportunities. 

Technology supporting the Change

The new agent lineup is supported by three major technologies, including:

• Long-horizon runtime: It is a new execution environment that lets an agent hold tasks or goals for days or weeks instead of a single chat. It combines memory, proper execution, and adjusts the plan as needed. Hunter is the first agent built on this, but Salesforce does say that they will be creating more agents adopting this technology. 
• Agent Script: It is the open-source language that defines agent behaviour to combine AI reasoning with rule-based logic. It is a great way to keep agents consistent with an organization’s existing rules and permissions. 
• Multi-agent orchestration: With the general availability of this technology, multiple agents get to work together. For example, a service agent can hand over tasks to the IT support agent without human support.  

Trusted Enterprise AI Harness & AI Control Plane

With new AI agents in Salesforce, the announcement also focuses on the governance layer, announcing the Trusted Enterprise AI harness and AI control plane. 

Trusted Enterprise AI harness 

The Enterprise AI harness is focused on building a trusted foundation around AI that brings together the ability to understand business, plan, reason, and take action while being in enterprise control. 

With the Enterprise AI harness, there are six major capabilities, including context, agency, action, governance, security, and models, which can be handled in one place. The technology focuses on delivering a common architecture built on customer relationships, processes, and controls. 

AI Control Plane

Alongside the AI harness, the AI control plane provides a unified platform to manage, see and control AI agents across the enterprise. Customers can use the new AI agents all together or just the ones they need with their existing Salesforce ecosystem and with third-party models, agents, and systems. 

The positioning of the AI agent and governance layer lines up with the broader shift in how enterprise buyers see AI. By naming agents around specific jobs, Salesforce is moving towards measurable, outcome-focused work. 

What does this mean for Salesforce Learners? 

With new AI agents and introduced capabilities, it also shapes the Salesforce career for both admins and developers. The major changes or shifts would be seen in the following ways. 

• Agent building is now a core skill: With Agentforce job-ready agents, the skill demand shifts from managing features like customization and design to learning agent scripting, prompt design, and multi-agent orchestration. 
• New Trailhead grounds: With new technologies, modules and certifications will expand toward long-horizon runtime, agent scripting, and AI control plane. Get a head start once enterprise adoption happens by getting hands-on understanding of the same. 
• Governance knowledge is now important: Understanding the trusted Enterprise AI harness and the six capabilities alongside it will be more important. Enterprises will start looking for admins with governance and compliance knowledge. 
• Customization skills are still valuable: As the seven agents can be renamed and adapted as per the organization’s process, the capability to customize agents will become a common and hireable skill. 
• Hands-on practice matters more than theory: With pre-built agents like Hunter or working on long-horizon runtime, still in rollout, learners who can set up sandbox pilots will be better positioned. 

WIth these, the question arises: Will AI replace Salesforce professionals? No, not the ones who understand the importance of upgrading and adopting new technologies with time. The new AI capabilities also bring opportunities and areas to function and learn, expanding new professional horizons.

Conclusion

Salesforce new AI agents 2026 introduce a clear shift in how AI is to be used and integrated into business processes. Whether Casey, Paige, Carter, Hunter, Marshall, Piper, and Fin will live up to the vendor’s expectations will have to be tested on a large scale. However, the governance layer and trusted enterprise AI harness is a larger shift once the AI agents live on. 

With Dreamforce 2026 new updates and releases, there are several AI upgrades expected in the days ahead. These changes not only impact business processes but also the overall learning for admins and developers, focusing more on AI capabilities and the governance layer. 

Author

Shrey Sharma

Shrey Sharma is the Founder of S2 Labs and a 2019 Salesforce MVP. He's trained 50,000+ students into Salesforce careers and runs Salesforce Hulk, the largest Salesforce-focused YouTube community. He's been a featured speaker at Salesforce community events worldwide, and his mission with S2 Labs remains simple: real mentorship, hands-on projects, and a clear path from classroom to career.