What are Salesforce's Agentforce named agents?

Salesforce launched seven named AI agents on 11 September 2026: Casey (customer service), Paige (IT/HR), Carter (commerce), Hunter (outbound sales, in pilot), Marshall (supply chain), Piper (inbound pipeline), and Fin (customer experience). Six are generally available; Hunter uses a new long-horizon runtime.

What is a long-horizon AI agent runtime?

A long-horizon runtime allows an AI agent to pursue a goal over days or weeks rather than completing a single-session task. Salesforce's Hunter agent uses this to track sales prospects across multiple touchpoints over time, maintaining context between conversations without being reset.

Should UK SMEs buy packaged AI agents from their software vendors?

Packaged agents like Agentforce integrate immediately with data already in that platform and need no developer to configure basics. The trade-off is scope: they cannot access data held in other tools. Businesses with operations spread across multiple platforms may need a platform-agnostic agent instead.

How widely will AI agents be embedded in business software by end-2026?

Gartner forecasts that 40% of enterprise applications will include task-specific AI agents by end-2026, up from less than 5% in early 2025. Global spending on agentic AI is projected to reach $201.9 billion in 2026 - a 141% increase on 2025, according to Gartner.

Salesforce Agentforce Launches Seven Named AI Agents: What UK SMEs Should Make of It

On 11 September 2026, Salesforce launched seven named AI agents - each with a job title, a defined business function, and months of operational memory - as part of its Agentforce platform ahead of Dreamforce 2026. For UK SMEs already using Salesforce or currently mapping out their AI agent strategy, the launch is a useful benchmark for how far the market has moved.

What are the seven Agentforce agents?

Each agent covers a specific business function:

Agent	Function
Casey	Customer service and help
Paige	IT support and HR requests
Carter	Commerce and shopper assistance
Hunter	Outbound sales pipeline (currently in pilot)
Marshall	Supply chain management
Piper	Inbound pipeline and lead qualification
Fin	Customer experience

Six agents are generally available. Hunter - the outbound sales agent - remains in pilot and is the first to use Salesforce's new long-horizon runtime. All seven run on Salesforce's Customer 360 data platform and operate within each company's existing security rules and permissions.

What is a long-horizon agent runtime and why does it matter?

Most AI agent interactions today are single-session: ask a question, get an answer, done. Hunter's long-horizon runtime is designed to pursue goals over days or weeks - tracking a prospect across multiple touchpoints, adjusting outreach based on what has and has not engaged, and maintaining a persistent objective without being reset.

For UK SMEs with a long B2B sales cycle - a typical professional services firm or a capital-equipment supplier - this persistent tracking is the practical value. The agent holds context on where each deal stands and acts on it between conversations, rather than only when a salesperson opens the CRM.

What does this launch signal for UK SMEs more broadly?

Gartner forecasts that 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from less than 5% at the start of 2025. The Agentforce launch is one high-profile example of a pattern appearing across all major business software categories: the software vendor embeds an agent so you do not have to build one.

The packaged approach has a clear appeal. Integration is immediate - the data is already in Salesforce, the workflow requires no external connector, and basic configuration does not need a developer. The trade-off is scope. An agent that only operates within Salesforce cannot see your accounting data in Xero, your project status in Asana, or your open tickets in Zendesk. For businesses whose operations span multiple platforms, packaged agents from individual vendors leave gaps.

A platform-agnostic agent, built to connect across your full tool stack, can coordinate across all of them. The design phase takes longer - you need to define what each workflow actually does - but the resulting automation is not bounded by a single vendor's data model.

Three questions to guide your decision

Packaged vendor agents and custom-built agents are not mutually exclusive. The right starting question is what your business actually runs on:

Is one platform central to your operations? If Salesforce is where your team works every day, Agentforce is worth evaluating on its own merits.
Is your data connected across tools? Agents only act on data they can reach. A siloed stack constrains any agent, packaged or custom.
Are you buying automation or buying an outcome? Define the outcome first - a qualified pipeline, a served customer, a resolved ticket - then identify the tool that delivers it.

The seven named Agentforce agents set a visible standard for what a job-ready AI agent now looks like in commercially available software. Whether your answer is Agentforce, a custom solution, or a mix of both, that standard is useful: each agent has a defined role, operates within your data and rules, and can be turned on or off without rewriting your whole tech stack.