The wait is over! At Dreamforce 2025, Salesforce officially unveiled Agentforce 360, its most advanced AI platform yet. According to Salesforce’s official newsroom update, the launch marks a major step toward the era of the “Agentic Enterprise.
While the discussion around “agentic AI” is still new for many technology firms, Salesforce is moving beyond concepts and delivering practical, goal-oriented solutions. Salesforce executives noted that they have spent the last year refining their agent technology through four major releases and thousands of customer deployments.
The core message from Marc Benioff, Chair and CEO, is clear: we are entering the “Age of the Agentic Enterprise.” This is a vision where AI agents don’t just assist—they actively work, collaborate, and take action alongside human employees. Agentforce 360 promises to be the single, trusted platform connecting humans, data, and intelligent agents.
But what makes this version different, and how can your organization use it to grow?
The Four Core Pillars of Agentforce 360: A Unified AI Platform
The power of Agentforce 360 comes from combining four key components. These four pillars ensure that AI agents have the foundation (Platform), the intelligence (Data), the purpose (Apps), and the workplace (Slack) they need to succeed.
1. Agentforce 360 Platform: The Foundation for Reliable Enterprise AI
This is the foundation where the agents are built and controlled. The new platform introduces several tools designed to make agents more dependable, accurate, and easy to govern.
2. Data 360 (Data Cloud): Grounding Agentforce 360 with Trusted Context
AI agents are only as smart as the data they use. Data 360 acts as the unified data layer, providing agents with essential context, whether the data is structured (like records) or unstructured (like documents).
3. Customer 360 Apps: Agents Embedded in Every Workflow (Sales, Service, Marketing)
This is where the business logic lives. Salesforce’s Customer 360 applications (Sales, Service, Marketing, etc.) are now fully embedded with agents that deeply understand specific business processes and can take action within those workflows.
4. Slack Integration: The Agentic OS and Collaboration Hub for Agentforce
Slack serves as the “Agentic OS for the Enterprise.” It is the central conversational interface where human teams and AI agents work together in real time, sharing knowledge, initiating actions, and tracking progress.
Key Innovations in Agentforce 360: Hybrid Reasoning and Governance
The Agentforce 360 Platform brings several groundbreaking features designed to give developers and business users maximum control and flexibility over agent behavior.
Hybrid Reasoning and Agent Script: Ensuring Predictable Agentforce Behavior
One of the main challenges in AI is balancing the creative nature of Large Language Models (LLMs) with the need for accurate, business-safe outcomes. Hybrid Reasoning solves this.
- Atlas Reasoning Engine: The core engine is now fully configurable. This allows teams to choose when an agent should use the flexible creativity of an LLM and when it must follow strict business logic.
- Agent Script: This is a major new component. Agent Script is a human-readable scripting language built specifically for controlling agents. It lets you define exact guardrails, conditional logic (if/then rules), and precise tool use. This ensures agents perform complex tasks predictably and safely.
- Model Flexibility: The platform now supports Google Gemini as a model for the Atlas Reasoning Engine, joining OpenAI and Anthropic, giving enterprises more choices for their agent backbone.
The Agentforce Builder: Low-Code Development for Conversational Agents
The days of complicated configuration are over. The new Agentforce Builder is a conversational development studio where teams can design, test, and deploy agents using natural language.
- Unified Workspace: Teams can draft, test, and deploy agents in a single interface, eliminating slow cycles.
- Rapid Iteration: It offers real-time simulation and live debugging, allowing developers to see the trace data for each step of an agent’s reasoning and make instant adjustments.
- Agentforce Vibes: This feature brings low-code development to agent creation, letting builders “vibe-code” apps grounded in company data and governance.
Agentforce Voice: Transforming IVR with Personalized, Real-Time Conversations
Agentforce 360 introduces a native voice layer that revolutionizes IVR (Interactive Voice Response) systems. This isn’t just text-to-speech; it’s a fully personalized, low-latency, real-time conversational agent.
- Brand Alignment: Agents can speak in a fully personalized voice aligned with the company’s brand, trained on their specific data and tone.
- Real-Time Responsiveness: The system is built for real-time interactions, executing actions like updating CRM records or triggering workflows while maintaining a near-human conversational flow.
- Fluid Conversations: Voice agents are interruptible, and every word is transcribed live in the Service Console, allowing human supervisors to take over instantly when necessary.
Data 360: How Unstructured Data Powers Agentforce 360 Accuracy
The term “Data Cloud” is now being unified under the name Data 360. This rebranding emphasizes its role as the source of truth that powers the agents.
Intelligent Context: Extracting Value from PDFs and Unstructured Data
Most enterprise knowledge sits outside of neat CRM fields, stored in PDFs, images, flowcharts, and long documents. Intelligent Context is the feature that changes this.
- Automated Extraction: It automatically converts this complex, unstructured content into structured, actionable data that Agentforce can use.
- Agent Grounding: This ensures that AI agents are grounded in the complete reality of the business, not just structured data, leading to more accurate and relevant answers.
- Low-Code Setup: Salesforce designed the pipeline so that document processing can be configured in hours, dramatically speeding up the deployment of context-aware agents.
Tableau Semantics: Standardizing Data Language Across the Agentic Enterprise
To ensure every agent, report, and employee uses the same definitions, Tableau Semantics translates complex data metrics into common business language using the Customer 360 Semantic Data Model (SDM). This feature ensures that terms like “qualified lead” mean the exact same thing across Sales, Marketing, and Service, no matter where the data is being accessed.
Also Read
Don’t forget to checkout: RAG 101: What It Is and Why It Matters.
Autonomous Action: Agentforce 360 in Sales, Service, and Commerce
The most exciting part of Agentforce 360 is how it transforms daily workflows across the entire enterprise. Agents are now embedded into every part of the Customer 360 suite, moving from simple assistance to autonomous action.
| Application | Agent Action Focus |
|---|---|
| Agentforce Sales | Automates prospecting, lead qualification, and quoting; recommends the next-best actions instantly. |
| Agentforce Service | Powers the Command Center for proactive, always-on support, managing cases end-to-end. |
| Agentforce Marketing | Autonomously builds, launches, and optimizes campaigns and customer journeys. |
| Agentforce Field Service | Automates scheduling and provides hands-free data capture through Voice to Form during service calls. |
| Agentforce IT Service | Reduces IT costs by providing 24/7 conversational resolutions, ending the traditional “portal-to-ticket era.” |
| Agentforce Commerce | Boosts shopper conversion with conversational, guided shopping experiences and checkout support. |
Furthermore, Salesforce is accelerating time-to-value with Agentforce 360 for Industries, offering pre-built solutions for complex fields like Life Sciences, Public Sector, and Manufacturing.
Slack: The New Control Center
Slack is positioned as the true Agentic OS for the Enterprise. It is no longer just a chat application but the unified space where humans, agents, apps, and data connect seamlessly.
The new features elevate Slack’s role:
- Reimagined Slackbot: The basic reminder tool is now a sophisticated, context-aware AI assistant. It can sift through conversations, summarize messages, take meeting notes in Huddles, and prioritize tasks based on the user’s role and data in Customer 360.
- Channel Expert Agent: This agent provides always-on, real-time expertise directly inside a Slack channel, grounded in enterprise knowledge.
- Model Context Protocol (MCP): This new protocol ensures open interoperability. It integrates third-party AI models and systems (like Anthropic and Dropbox) into Agentforce, making Slack the central hub for all agent activity, regardless of the underlying technology.
The Road to Dreamforce 2025: Agentforce Development Timeline
The release of Agentforce 360 is the result of continuous development over the past year. This is not a starting point, but the next major leap in a clear roadmap:
- October 2024: Initial launch of Agentforce.
- December 2024: Agentforce 2, focusing on improving the Atlas Reasoning Engine for more grounded results.
- March 2025: Agentforce 2dx, enabling agents to be embedded in proactive, triggered, and cross-functional workflows.
- June 2025: Agentforce 3, preparing for enterprise scale with enhanced governance and interoperability.
This history shows that the new Agentforce 360 is a mature, tested platform built on real-world customer deployments.
Key Takeaways from Dreamforce 2025
For quick reference, here are the most important points about the Agentforce 360 launch at Dreamforce 2025:
- Autonomous Action: Agentforce 360 is Salesforce’s flagship autonomous AI platform, designed to enable agents to reason, act, and collaborate without constant human oversight.
- Four Pillars of Power: The platform is built on four unified components: Agentforce 360 Platform (The Engine), Data 360 (The Context), Customer 360 Apps (The Action), and Slack (The Workplace).
- Predictable AI: New Hybrid Reasoning combined with Agent Script gives developers deterministic control over agent behavior, ensuring accuracy and compliance for complex business tasks.
- Unstructured Data Mastery: The new Intelligent Context feature allows agents to automatically extract, structure, and use knowledge from PDFs, images, and other unstructured content, grounding them in the complete institutional memory of the business.
- Voice Revolution: Agentforce Voice transforms customer service by providing fully personalized, low-latency, and fluid conversational agents integrated directly into the platform.
- Unified Workspace: Slack is now the “Agentic OS,” serving as the central hub where human employees and AI agents seamlessly collaborate and orchestrate workflows across all business functions.
Conclusion: Why Agentforce 360 Matters for Your Business Growth
Dreamforce 2025 confirms that the “Age of the Agentic Enterprise” is not future speculation—it is here. The unveiling of Agentforce 360 provides a unified, highly governed platform that connects data, collaboration, and CRM with autonomous AI agents.
This marks a significant shift, empowering every employee to move from simple data entry and assistance to strategic, high-value work. The platform is available globally today, ready to redefine how businesses operate.
