Agentforce Coworker (sometimes searched as "Salesforce Coworker") is Salesforce's new autonomous AI teammate that lives inside your CRM. Announced by Marc Benioff in May 2026, it puts AI in every Salesforce search bar so you can ask questions in plain language, surface insights, and trigger work directly from Global Search.
This guide explains what Agentforce Coworker is, its core capabilities, practical use cases, and how to turn it on — written for mid-market Salesforce admins and business leaders deciding whether to adopt early.
Important status note: Agentforce Coworker is currently in Beta (Salesforce's setup is documented as "Turn On Agentforce Coworker (Beta)"). Features and availability may change before general availability. Always verify the current state in official Salesforce documentation before planning a rollout.
Agentforce Coworker is an autonomous AI teammate built into Salesforce that acts as a single entry point to every agent your team creates. It reimagines Salesforce search — letting you converse with your CRM in plain language from Global Search — and it can discover information, analyze data, and take action on your behalf. It matters for mid-market admins and business leaders who want a unified, governed front door to Agentforce rather than a sprawl of disconnected agents. Because Coworker indexes across many sources and takes action, access control and data governance are central to using it safely. Vantage Point helps teams enable Agentforce Coworker on a trustworthy data foundation.
Agentforce Coworker is an autonomous AI teammate inside Salesforce that serves as the single entry point to every agent your team creates. According to Salesforce's Agentforce Coworker product page, as you add more agents, Coworker evolves and automatically knows which agent to use for a given task.
Its defining idea is "AI in every search bar." Instead of clicking through reports and list views, you converse with your CRM in plain language directly from Salesforce Global Search. Coworker knows your business context from day one — opportunities, cases, accounts, pipeline, and customer history — and can take action in real time.
The name causes some confusion. David and others have referred to it as "Salesforce Coworker," but the official product name is Agentforce Coworker. We use the official name throughout this guide.
Most organizations adopting Agentforce end up with multiple agents — a service agent, a sales agent, an internal operations agent — and no consistent way for employees to reach them. Agentforce Coworker solves that by acting as a unified front door: one place to ask, analyze, and act, with Coworker routing to the right agent automatically.
For mid-market teams, the practical reality is that data is often messy or disconnected across systems. Coworker is most useful when it sits on connected, trustworthy data. That makes data governance and a sound data foundation a strategic prerequisite, not an afterthought — a point we explore in why most AI projects fail without a data foundation.
Salesforce describes three core capability pillars on the official product page. The table below summarizes each.
| Capability Pillar | What Salesforce Says It Does | Practical Example |
|---|---|---|
| Discover anything | Connects and indexes across 300+ enterprise sources to surface the right answers with full business context. | Ask "What's the latest on the Acme renewal?" and get a context-aware answer pulled from CRM and connected sources. |
| Analyze and explore insights | Connects across all your enterprise data to surface proactive insights the moment you need them, getting smarter as your business evolves. | Surface at-risk pipeline or unusual case trends without building a report. |
| Plan and act | Triggers flows, activates agents, and keeps work moving autonomously so nothing stalls while you're heads-down or away. | Kick off a flow or hand a task to an existing agent straight from search. |
The throughline is that Coworker does not just answer questions — it connects to your data, reasons over it, and can act. That action-taking ability is what makes governance so important.
Salesforce reports that setup is largely automated. One customer, Andrew Russo, VP of Business Systems at BACA Systems, described it as "the easiest thing I've ever setup for our entire Salesforce org." Use this checklist as a starting point, and follow the official Salesforce Developers turn-on guide for current, authoritative steps.
Because this is a Beta feature, treat any specifics — availability, limits, configuration screens — as subject to change and verify them in Salesforce's documentation.
Frame these as practical examples of what Coworker is designed to do, not guarantees:
Agentforce Coworker indexes across many sources and takes action. That changes the stakes. An AI teammate that can trigger flows and activate agents needs clear access controls, accurate data, and a defensible governance model — otherwise it can act confidently on bad or unauthorized information.
You do not need perfect data to start. But you do need a deliberate approach to data quality, connectivity, and permissions. This is where a data foundation and integration strategy pays off, and it connects directly to broader work like Salesforce Data Foundations and MDM. For a fuller view of how Coworker fits the wider agent portfolio, see the Salesforce Agentforce ecosystem guide.
Vantage Point is a US-based, employee-owned, senior-only Salesforce partner focused on the mid-market. We are an Anthropic Registered partner and use our VALUE Methodology to deliver outcomes, not just configuration.
For Agentforce Coworker, we help teams enable it on a trustworthy footing: assessing data readiness, designing access controls and governance for action-taking AI, and connecting the data sources Coworker depends on. Our Salesforce implementation and advisory team pairs with our AI and analytics practice so you adopt Agentforce Coworker with confidence.
Ready to evaluate Agentforce Coworker? Talk to Vantage Point about an Agentforce Coworker enablement and data-readiness consultation. We will help you decide whether to adopt the Beta now and how to do it safely.
Agentforce Coworker is Salesforce's autonomous AI teammate built into the CRM. It acts as a single entry point to every Agentforce agent your team creates and lets you search, analyze, and take action in plain language from Salesforce Global Search.
Yes. Some people search for "Salesforce Coworker," but the official product name is Agentforce Coworker. They refer to the same Salesforce AI teammate.
Salesforce provides an automated Agentforce Coworker (Beta) setup that provisions much of the required infrastructure. Confirm your org's eligibility and Beta availability, validate connected data sources, set access controls, then follow the official Salesforce Developers turn-on guide for current steps.
No. As of this writing, Agentforce Coworker is in Beta, so features and availability may change. Check current Salesforce documentation for the latest status, editions, and any limits.
Per Salesforce, it has three core capabilities: discover information across 300+ connected enterprise sources, analyze data to surface proactive insights, and plan and act by triggering flows and activating agents. It knows your business context — opportunities, cases, accounts, and pipeline — from day one.
It works best on connected, trustworthy data because it indexes across many sources and takes action on your behalf. You do not need perfect data to start, but accurate data, clear connectivity, and strong access controls make it far more useful and safer.
Salesforce has not published pricing details we can cite here, and Agentforce consumption is generally tied to Salesforce's Flex Credits model. Confirm current pricing in official Salesforce documentation, and review how Flex Credits work to plan for consumption.
It depends on data readiness and governance maturity. If your CRM data is reasonably connected and you can define who uses Coworker and what it can do, an early pilot can be valuable. Vantage Point can assess your readiness and help you adopt it safely.