Table of Contents
Salesforce Dreamforce 2025 has just concluded with a brand new launch: Agentforce 360. After spending a whole year releasing four major updates in the Agentic AI scene, Salesforce has yet again upgraded its flagship AI product. This time, the upgrade is not just a newer version but a complete solution that will lead businesses to become what Salesforce claims to be, an ‘Agentic Enterprise’.
“Agentforce 360 connects humans, agents, and data on one trusted platform, helping every employee and every company achieve more than they ever thought possible,” said CEO Marc Benioff in his opening keynote of the annual event.
Let’s dig deeper into what this new suite of Salesforce Agentforce is and how it is better than the previous versions.
What is Salesforce Agentforce 360?
Utilising the core Salesforce Trust Layer and built upon the previous Agentforce platform, Agentforce 360 is an amalgamation of four core features. It includes Agentforce Builder for faster agent development, Agent Script for greater control of agent behaviour, Agentforce Voice for natural on-brand conversations, and Intelligent Context to ground agents in your complex unstructured data.
Here is a recap of Agentforce’s evolution in a year:
- Agentforce (October 2024): Salesforce brought the first enterprise AI agent platform to the world.
- Agentforce 2.0 (December 2024): Major improvements were introduced in the Atlas Reasoning Engine for more predictable, grounded results.
- Agentforce 2dx (March 2025): A new developer-friendly version that enables agents to be embedded in any workflow.
- Agentforce 3 (June 2025): This version was meant to enhance interoperability and governance, preparing enterprises for scale.
The new Agentforce 360 platform is an extension of these upgrades and claims to transform Salesforce beyond a CRM that serves customers. It aims to empower employees working on it and pave the way for Agentic Enterprise.
The Platforms Behind the Shift
Salesforce brings together four products to form a single ecosystem, which can be called the Salesforce AI platform. Each one serves a unique purpose to provide an enhanced experience to both users and employees collectively.
- The foundation is now the Agentforce 360, which provides enterprise-grade AI agents. The renewed platform features conversational builder and hybrid reasoning for increased control and accuracy alongside voice capabilities.
- The second pillar of agentic enterprise is going to be Salesforce Data 360 (previously Data Cloud). It includes the power of Intelligent Context and Tableau Semantics to convert data into context for AI.
- Besides these, the Customer 360 Apps of Salesforce are going to provide business logic for every enterprise. This includes memory saved and actively being gathered about selling, marketing, operations, etc.
- Lastly, it uses the Slack platform for a seamlessly integrated conversational interface where humans and agents can work together.
Let’s dive deeper into the platform and its components that contribute to a faster, better Salesforce AI platform.
What is Inside the Agentforce 360 Platform?
Like we mentioned before, the Agentforce 360 Platform is made up of four components that shape the way it works.
Agentforce Builder
Agentforce Builder serves as the single conversational space to not only build reliable agents but also test and refine them. This eliminates the need to switch between different development cycles to bring everything together in one place. The vibe-coding technique can be helpful here to iterate faster, safer, and more collaboratively. Every agent compiles into a portable JSON file, making versioning and sharing a breeze.
Key capabilities include:
- Unified workspace: Instead of the conventional build-test loop, combine everything, including deployment, into a single interface.
- Multi-view editing: Easily transition between a script view to fit any workflow and a doc-like editor with autocomplete on a low-code canvas.
- Quick iteration using live debugging and simulation: See trace data for every step of the reasoning process, preview actual user interactions in real time, and make quick adjustments.
- Governance at the enterprise level: based on Salesforce’s privacy and security features to guarantee safe deployment and compliance.
- Agentforce Vibes: Agentforce Vibes is an AI coding companion that works with you like a pair programmer, comprehending the context of your Salesforce project and carrying out tasks for you to speed up development.
Agent Script
Agent Script blends the predictability of code with the inventiveness of AI to ensure that agents always behave exactly how you want them to. Conditional logic, accurate tool use, and guided, deterministic controls are all made possible by Agent Script. It also gives you the ability to specify agent behaviour using natural language. Agentforce can confidently take on even your most challenging jobs because of this increased level of control.
Key Capabilities include:
- More predictable agents: Teams may precisely design complicated agent logic, such as conditionals, “if/then” rules, and hand-offs, using this portable, human-readable JSON expression language.
- Hybrid Reasoning: The Atlas Reasoning Engine’s Hybrid Reasoning feature allows teams to blend the innate creativity of LLMs with the certainty of structured business logic, resulting in more dependable and predictable agent behaviour.
Agentforce Voice
Agentforce Voice provides organic, human-like dialogues that easily integrate with all aspects of your company and capture the distinct tone and personality of your brand. It turns every call into a dynamic, real-time engagement that is wholly integrated into every Salesforce application, since 80% of all enquiries begin over the phone. For dependable, enterprise-ready performance, each agent is exposed to thousands of real-world circumstances, such as busy callers, noisy backgrounds, and intricate, multi-step procedures.
Key capabilities include:
- Fully customised voice agents: Trained on the company’s data, tone, and contextual knowledge, each AI agent speaks in a voice consistent with the brand.
- Low-latency, ultra-realistic interactions: Designed for real-time responsiveness, Agentforce Voice accomplishes conversational flow that is almost human while carrying out tasks like calling APIs, updating CRM records, or starting workflows.
- Smooth integration: With first-rate support for Amazon Connect, Five9, NiCE, and Vonage, Agentforce Voice is a contact centre stack extension that is natively linked to Salesforce.
- Conversations that are fluid and interruptible: Salesforce Voice transcribes every word in real time, providing human agents with a real-time view from the agent dashboard to enable quick, smooth takeovers when necessary.
- Voice-first platform architecture: Enhanced latency, speech realism, and deep language understanding make it not just a new modality – but a complete platform experience built to meet customers where they are.
Intelligent Context
Business-specific context is the first step towards enterprise-grade agent experiences. Your AI agents will be firmly rooted in the intricate, unstructured data of your company thanks to Intelligent Context. This feature, which is powered by the new Data 360, automatically extracts, formats, and displays this data so that your agents can provide precise, pertinent, and business-specific responses. You can set up document processing in hours rather than days with a low-code pipeline. Additionally, context can be configured to allow agents to interpret the same content from multiple business perspectives. As a result, your teams will have rich, contextual intelligence in every encounter thanks to agents that can reason with the appropriate facts.
Key capabilities include:
- Automated context extraction: Provides Agentforce with organised, actionable data from complicated, unstructured texts.
- Configuration in low code: Without requiring a lot of engineering work, quickly set up unstructured data pipelines in a matter of hours rather than days.
- Multimodal support: Manages a variety of content formats, including flowcharts, tables, PDFs, and photos.
- Integration with Data 360 is smooth: Use AI to build a search index, then implement it globally for safe, contextual, and enterprise-ready operations.
Other than these four components, the Agentforce 360 will utilise the existing application base of Salesforce.
Customer 360 Apps
Salesforce apps now include conversational agents that can operate directly in business operations.
- Agentforce Sales: This uses next-best actions to automate prospecting, qualification, and quotation.
- Agentforce Marketing: This creates and starts assets, campaigns, and trips on its own.
- Agentforce Service: Providing proactive, round-the-clock support by powering the Command Centre for Service.
- Agentforce Field Service: Provides hands-free data collection using Voice to Form, integrates live maps through Esri, and automates scheduling.
- Agentforce Revenue Management: With the help of AI agents that enable sales, operations, and billing teams with agentic quoting, agentic billing, and consumption management, it generates steady growth.
- Agentforce Commerce: With conversational, guided shopping experiences and checkout, it helps raise cart size and boost consumer conversion.
- Agentforce IT Service: By putting an end to the “portal-to-ticket era” and deploying specialised AI agents to provide round-the-clock conversational remedies wherever employees work, it helps lower IT costs and increase employee productivity.
- Agentforce 360 for Industries: With new pre-built, sector-specific solutions like Agentforce Life Sciences, Agentforce Public Sector, and Agentforce Manufacturing, it offers a quicker time to value.
Slack: The ‘Agentic OS for the Enterprise’
Slack has evolved into a system where humans, agents, apps, and data connect in real time.
- Slack-First Apps: Tableau, IT Service, HR Service, and Agentforce Sales. Next, reveal insights and finish tasks without ever leaving Slack.
- Channel Expert Agent: Real-time enterprise knowledge search-powered, always-on channel expertise.
- Enterprise Search: Natural language responses for Jira, GitHub, Google Drive, and other platforms.
- Slackbot has been redesigned as a context-aware helper for creating huddle notes, message summaries, and assistance.
- Model Context Protocol (MCP): Uses new APIs to integrate third-party AI for automation and real-time knowledge, such as Anthropic, Dropbox, and OpenAI.
In Summary
Salesforce is heading towards creating an Agentic Enterprise, a universal experience for all businesses. Being customer zero and testing Agentforce 360 to improve their employee coordination conveys that AI is here to elevate your life. Agentforce 360 is available globally today, with continued innovation arriving in pilot and beta over the coming months. The evolved Agentforce platform is going to ease business processes with enhanced data ingestion through Data 360 and voice coordination through Agentforce voice. To keep up with the platform and its changes, you must learn and upskill about Agentforce while there is still time.
