Salesforce TDX 25: Key Highlights & Announcements
Salesforce TrailblazerDX (TDX) 2025 has started, bringing a wave of exciting announcements, with AgentForce taking center stage as the biggest highlight of the event. This next-gen AI automation platform just got smarter, faster, and more developer-friendly, making it easier than ever to build AI agents that reason, take action, and seamlessly integrate into workflows.
If you missed the keynote, don’t worry—we’ve got you covered. This blog highlights all the major announcements from TDX 2025, breaking down everything you need to know in one place.
#1. Data Cloud & AgentForce Are Now Live in Developer Edition
One of the biggest announcements at TDX 2025 is the availability of AgentForce and Data Cloud in Developer Edition. This is one of the major announcements, as it allows you to experiment, build, and deploy AI agents without limitations. Unlike Dreamforce scratch orgs that expired in 9 hours, this new Developer Edition never expires—as long as you log in every 45 days.
With this Developer Edition, Salesforce is giving you the access to:
150 LLM outputs per hour
1 Data Space
10GB of Storage
Agentforce Developer Edition | Source: Salesforce
#2. AgentForce 2DX – Smarter, More Proactive AI for Everyone
Just last year, in December 2024, Salesforce launched AgentForce 2.0, an upgraded version of its AI-powered digital labor platform. Now, at TDX 2025, Salesforce has taken it even further by introducing AgentForce 2DX—an enhanced version with greater control, deeper integrations, and proactive automation capabilities.
With this upgrade, AI-powered automation is no longer just reactive—it’s proactive, secure, and action-driven. Let’s take a look at what’s new.
AgentForce 2DX | Source: Salesforce
1. Variables (GA March 25)
The all new Variables feature allows AI agents to work with deterministic logic instead of relying solely on general AI reasoning. This means AI agents only access relevant data at the right time, ensuring better security and compliance.
For example:
A customer support agent using an AI-powered assistant will only see case-specific data for the current interaction rather than accessing broader company information.
An AI agent assisting HR or finance teams will retrieve only relevant session-based data, preventing unauthorized access to sensitive employee or financial records.
Developers can configure these variables using Agent Builder, defining filters to ensure AI agents access and process data only when required.
Variables in Agent Builder | Source: Salesforce
2. Tableau Next (GA Now)
AI responses have evolved beyond just text. With Tableau Next, AgentForce 2DX can now generate real-time visual insights, making it easier for employees, developers, and decision-makers to analyze complex data at a glance.
Instead of simply displaying numbers or summaries, AI agents can now create interactive charts, trend graphs, and dashboards—helping teams quickly understand key performance indicators, patterns, and anomalies.
3. Agentic API & Flow Triggers (GA now)
AI agents are no longer just reactive assistants responding to queries. Thanks to Agentic APIs and Flow Triggers, AI can now automate workflows, take proactive actions, and execute business processes in real time.
Traditionally, AI agents required human input to function. Now, with Flow Triggers, AI agents can run in the background, monitor business processes, and trigger workflows automatically when specific conditions are met.
#3.AgentForce Development Lifecycle
Salesforce has introduced new enhancements to the AgentForce development lifecycle, streamlining the process of building, testing, deploying, and observing agents for both low-code and pro-code developers. Let’s break it down:
AgentForce 2dx Development Lifecycle | Source: Salesforce
No-Code Development
1. Build Agents with AI Assistance
Creating an agent from scratch is now easier than ever. Previously, developers had to write detailed descriptions and define workflows manually. With the new AI-assisted feature, you simply describe what you want the agent to do, and the AI automatically generates the logic, topics, and actions. This no-code approach makes agent creation faster and more accessible, even for users without technical expertise.
Want to know more about AgentForce? Check out our blog on how AgentForce works.
AI generated Topics in Agentforce | Source: Salesforce
2. Adaptive Response Formats
AgentForce now supports Adaptive Response Formats, allowing you to customize how AI-generated responses are displayed. Instead of plain text, responses can be presented in interactive formats such as rotating cards, forms, or other visually engaging designs.
Adaptive Response Formats in Agent builder | Source: Salesforce
Pro-Code Agentforce Development
While the no-code approach simplifies agent creation, Salesforce has also introduced powerful tools for pro coders:
1. AgentForce CLI (GA April 2025)
Developers can now create and manage agents directly from the command line using Salesforce CLI. Integrated with tools like VS Code, this feature allows developers to retrieve or pull agent metadata from their org, making development more efficient and streamlined.
Agentforce CLI | Source: Salesforce
2. LWC in Responses (Developer Preview)
Similar to the Adaptive Response Format for no-code users, developers can build their own Lightning Web Components (LWC) to display agent responses exactly as they want.
Testing center (GA now)
Salesforce has truly done a great job by providing ways to build agents using both pro-code and no-code approaches, but what about testing?
At TDX, Salesforce announced the Testing Center, where you can test agents regardless of how they were created—whether through no-code tools or custom code. With the integration of low-code and pro-code environments, developers can now test agents seamlessly in the Testing Center, accessible via Salesforce Setup → Einstein Generative AI.
To run a test:
Name the test
Specify the agent being tested
Provide a description
Upload a CSV file with test data or generate test data using AI by describing your requirements and specifying the number of test cases needed.
Test agents with testing center | Source: Salesforce
Additional Enhancements in AgentForce Development
Salesforce has introduced several new tools and features to enhance the AgentForce development lifecycle, making it easier for developers to build, test, deploy, and observe agents. Here are the key enhancements:
1. DX Inspector
The most anticipated feature has finally been announced—DX Inspector. It provides a comprehensive view of all agent metadata, including flows, prompts, topics, and more. Additionally, it includes a change tracker, allowing developers to monitor and manage changes in real-time. This ensures transparency and control over the agent’s development lifecycle, making it easier to debug and optimize.
DX Inspector and change tracker | Source: Salesforce
2. AgentForce in DevOps Center (GA Now)
Deploying agents is now seamless with AgentForce in DevOps Center. This integration simplifies the deployment process, ensuring smooth transitions from development to production environments. Developers can promote agents through the pipeline, ensuring the right checks and balances are in place.
AgentForce in DevOps Center | Source: Salesforce
3. AgentForce Interaction Explorer (GA Fall 2025)
The upcoming Interaction Explorer will provide deeper insights into agent interactions, helping developers analyze and optimize performance for better outcomes. This tool allows you to explore conversations, identify patterns, and improve agent responses.
AgentForce Interaction Explorer | Source: Salesforce
4. Data Cloud Governance (GA Now)
With Data Cloud Governance, developers can ensure that agents operate within secure and compliant data environments. This feature allows you to set up rules to automatically tag and detect sensitive information across structured and unstructured data. Fine-grained access controls ensure that data is only accessible to authorized users, maintaining trust and reliability.
Data Cloud Governance | Source: Salesforce
#4. AgentExchange – The All New Marketplace for AI Agents
Salesforce saved one of its most revolutionary AI announcements for the end, and when Christophe Coenraets (SVP, Developer & Admin Relations, Salesforce) took the stage to unveil AgentExchange, it was nothing short of a historic moment for AI agents.
So, what is AgentExchange? Think of it as the AppExchange for AI agents—a platform where you can find prebuilt solutions for every industry, deploy them instantly, and even share or sell your own AI creations.
AgentExchange is designed to make AI agent adoption faster, simpler, and more scalable. Whether you’re a business looking to integrate AI into your workflows or a developer building custom agents, this marketplace has something for everyone.
Salesforce AgentExchange | Source: Salesforce
How to Get Started with AgentExchange
Simply head over to agentexchange.salesforce.com and explore a growing library of AI solutions tailored to your needs. And here’s the best part—Salesforce is planning to integrate AgentExchange directly into Agent Builder. This means you’ll soon be able to discover and extend your agents’ capabilities as you build them, all within the same workflow.
AgentExchange in Agent Builder | Source: Salesforce
What’s in It for Developers?
Remember the AgentForce CLI we mentioned earlier? Developers can use it to build custom agents, and if they decide to share or sell their creations, they can package the metadata and list it on the AgentExchange marketplace, providing global reach to your solution.
Conclusion:
Salesforce TDX 2025 has taken the agentic revolution to new heights, redefining the boundaries of AI automation. With groundbreaking updates like AgentForce 2DX, AgentExchange, and enhanced development tools, Salesforce continues to push the envelope, proving that innovation doesn’t stop at what’s already working—it evolves, refines, and reshapes the future of AI-powered workflows.
Stay tuned with us as we keep a close eye on Salesforce’s advancements, bringing you the latest updates and insights into the ever-evolving world of AI.
Let’s Talk!
Drop us a note, we’re happy to take the conversation forward
