Salesforce is rewriting the AI development playbook. On May 21, 2025, the company launched Salesforce Agentforce DX, a pro-code suite of tools that empowers developers to build, test, and manage AI agents, all from their local environment or IDE.
If you’ve been stuck bouncing between your IDE and Salesforce’s web UI, this is the update you’ve been waiting for. Salesforce Agentforce DX gives full control back to the developer, enabling local, source-controlled, and automated workflows that integrate seamlessly into modern DevOps.
Salesforce Agentforce DX Guide
What Is Salesforce Agentforce DX?
Agentforce DX is Salesforce’s new pro-code framework designed to give developers everything they need to build AI agents inside the tools they already use: Salesforce CLI, Visual Studio Code, and Code Builder.
It addresses one of the biggest pain points in AI development on Salesforce: context switching. Previously, building or testing agents required frequent jumps between the command line and the Salesforce UI. Now, you can stay in your IDE from start to finish.
Agentforce DX is part of the broader Agentforce platform, which includes Agentforce 2dx for proactive AI agents and the AgentExchange marketplace.
What’s Included in Agentforce DX?
Agentforce DX introduces a powerful set of tools that bring AI development into the modern software development lifecycle (SDLC):
1. Salesforce CLI Plugin for Agents
The new CLI plugin (sf agent) supports:
- Agent specification generation
- Agent creation and configuration
- Previewing agent behavior
- Running automated tests
No browser tab needed, just type and deploy directly from your terminal.
2. IDE Integration – VS Code & Code Builder
Agentforce DX includes a new VS Code extension (also compatible with Code Builder), offering:
- Real-time preview and testing
- YAML-based editing with full autocomplete
- No context switching from your dev workflow
Developers can now build AI experiences in the same place they write Apex or Lightning Web Components.
3. YAML Configuration for AI Agents
All agent metadata, behavior flows, and test scripts are stored in human-readable YAML. Salesforce also allows conversion from traditional XML metadata to YAML, giving pro-code developers a cleaner, more flexible way to version and review code.
Key Benefits of Agentforce DX
Smarter Development with Less Friction
Agentforce DX is purpose-built to make AI agent development feel native to developers. No need to learn new UI tools. Just stay in your IDE and write code.
Git-Friendly Metadata
YAML means better source control, better code reviews, and easier merge conflict resolution—all essential for scalable Salesforce DevOps.
CI/CD Ready
You can now include agent specs, tests, and behaviors in your CI/CD pipelines. This opens the door to test automation, performance monitoring, and rollback support—something previously difficult with point-and-click development.
Agent Simulation in Local Environments
Simulate agent conversations and test business logic against Apex classes—all in your local environment. This leads to faster debugging and fewer production errors.
Introducing Agentforce 2dx
Alongside Agentforce DX, Salesforce has launched Agentforce 2dx—a major leap in autonomous AI capabilities.
What’s New?
- Proactive agent behavior: Agents can now act autonomously based on data or system events.
- Multi-channel engagement: Agents operate across UI, mobile, chat, and voice platforms.
- Context awareness: AI assistants retain and use context to make better decisions over time.
- Autonomous decisioning: Agents can take actions like approvals, recommendations, and follow-ups without human intervention.
Agentforce 2dx is Salesforce’s answer to generative AI in enterprise automation—and Agentforce DX gives developers the tools to build it.
AgentExchange – The Marketplace for AI Agent Components
Salesforce also announced AgentExchange, a new marketplace for AI agent components, including:
- Agent behaviors
- Pre-trained conversation topics
- Third-party integrations
Launch partners include Google Cloud, DocuSign, and Box—each contributing building blocks that developers can drag-and-drop or customize in YAML.
Developers can also monetize their own agent components, creating a shared ecosystem for intelligent enterprise tools.
Getting Started with Agentforce DX
To start building AI agents with Agentforce DX:
1. Update Salesforce CLI
sf update
2. Install the Agent Plugin
sf plugins install @salesforce/plugin-agent
3. Create Your First Agent
sf agent generate agent-spec
4. Install VS Code Extension
Search for Agentforce DX in VS Code Marketplace or Open VSX.
5. Use Salesforce’s Free Developer Environment
Salesforce offers a no-cost, full-featured Dev Hub that includes Agentforce DX, Data Cloud, and Einstein AI.
Success Stories of Salesforce Agentforce DX
The Adecco Group
Leveraged Agentforce to match candidates faster and streamline hiring workflows with autonomous agents.
OpenTable
Created intelligent customer service agents that handle reservations and upsell premium dining experiences in real time.
Vivint
Used AI agents to handle 60% of customer inquiries autonomously, improving agent efficiency and NPS scores.
Why Agentforce DX Matters for Developers
Salesforce has long supported low-code/no-code AI, but pro-code developers were often left out of the conversation.
With Agentforce DX, Salesforce is:
- Bringing AI to DevOps
- Supporting enterprise-scale versioning
- Making agent development testable, repeatable, and collaborative
Agentforce DX empowers developers to treat agents like code—not like configuration. That’s a huge shift in AI development.
Final Thoughts
Agentforce DX isn’t just a tooling update—it’s a developer revolution inside the Salesforce ecosystem.
For the first time, Salesforce has fully embraced the pro-code paradigm for AI. If you’re a developer looking to:
- Stay inside your IDE
- Own your AI agents like code
- Integrate AI into your SDLC
Then Agentforce DX is built for you.
Now is the time to rethink how you develop intelligent assistants on Salesforce—and start building smarter, faster, and better AI-powered solutions.
FAQs
1. What is Agentforce DX?
A pro-code toolset by Salesforce that lets developers build, test, and manage AI agents using CLI and IDEs—no UI needed.
2. Why is Agentforce DX a big deal?
It finally brings AI agent development into the developer’s IDE, supports YAML, Git, and CI/CD, and removes the need to use the Salesforce UI.
3. What tools does it work with?
Salesforce CLI, Visual Studio Code, and Code Builder.
4. What is Agentforce 2dx?
A next-gen upgrade enabling proactive, autonomous AI agents that can act on data and run in the background.
5. Can I test agents locally?
Yes! Use the CLI or VS Code extension to simulate agent behavior without pushing to an org.
6. Is YAML used in Agentforce DX?
Absolutely. Agent specs, test data, and configs are stored in clean, readable YAML for easy editing and versioning.
7. What is AgentExchange?
A new Salesforce AI marketplace where you can find or publish agent components like templates, actions, and integrations.
8. Is there a free version?
Yes, Salesforce offers a free dev org with Agentforce DX, Data Cloud, and tools preloaded.
9. Where can I learn more?
Check the Agentforce DX Developer Guide and the AgentExchange Marketplace.
