TDX (TrailblazerDX), where DX stands for Developer's Experience, is Salesforce's conference designed for developers, admins, architects, and partners to learn about and build with Salesforce solutions.
This year's event was entirely focused on Agentforce and Digital Labor (R)Evolution.
The 4th Industrial Revolution is here!
Keynote Highlights
AI's Inevitable Rise & Salesforce's AI-First Future
The keynote delivered some (not-so) surprising news—AI is here to stay, whether we embrace it or not. Salesforce is rapidly evolving into an AI-first company, making bold strides in Digital Labor.
A major announcement was Agentforce 2dx, with new features set to roll out in the coming months. Expect even more Agentforce updates at the upcoming Dreamforce Conference in October 2025.
Why the Rapid AI Shift?
Declining Populations Are Fueling the Need for AI. With global population growth slowing, businesses face potential stagnation. AI presents a solution—addressing labor shortages, sustaining economic growth, and enhancing productivity—while still valuing the human element.
Keep reading to discover all the latest updates to Salesforce products.
New Technology, New Mindset - How Salesforce integrated Agentforce into help.salesforce.com
Salesforce has completely overhauled its help portal, integrating Agentforce to streamline self-service. The new design features a much simpler UI and improved site navigation. Now, you can ask an agent (in natural language) any question, and it will provide answers along with relevant content links.
Shifting to an Agentic Portal Experience
This transformation requires a fundamental mindset shift from:
-
Navigation → Conversation
-
Content Management → Knowledge Engineering
-
Siloed Specialists → Cross-Functional Coalitions
-
User Interface → Flexible Engagement
-
Building Features → Designing Outcomes
Check out the screenshots below! If you remember the previous cluttered help portal, you're in for a pleasant surprise.
Building an Effective Agent
Now that your mindset has shifted toward agentic AI, the next step is figuring out how to build an effective agent.
The Four Key Steps:
-
Design – Define the agent’s purpose and capabilities.
-
Build & Test – Develop specialized agents and rigorously test them.
-
Deploy – Launch the agent and gather user feedback.
-
Govern – Continuously monitor, refine, and improve.
Step 1: Design the Agent
Document use cases and jobs to be done. Be as detailed as possible. Every agent has five key attributes:
-
Role – What job does the agent perform?
-
Data – What knowledge can it access?
-
Actions – What capabilities does it have?
-
Guardrails – What should it NOT do?
-
Channel – Where does it operate?
Step 2: Build & Test
Develop task-specific, specialized agents for each department or use case identified in the design phase. Thorough testing is crucial before deployment. You will learn more about testing in the Agentforce Testing Center below.
Step 3: Deploy & Gather Feedback
Roll out the agent and encourage users to provide feedback to improve performance.
Step 4: Govern & Refine
Deployment is just the beginning. Agents require a continuous refinement cycle:
Ideate → Configure → Test → Deploy → Supervise.
Regular monitoring, security & compliance audits, and user-driven refinements ensure long-term success.
Want to dive deeper into building effective agents?
Check out this blog post by Elements.cloud, one of the TDX presenters. They’ve developed a tool called Agent Designer to help streamline the process of designing and building agents.
Agentforce Testing Center
In the previous section, I highlighted testing as a key part of building and governing agents. Until now, testing agents in Salesforce has been a manual task—you’d open the agent in Agent Builder, start a conversation, and review the interaction logs.
But how do you scale that and test across multiple scenarios efficiently?
Enter the Agentforce Testing Center — a first-of-its-kind AI agent lifecycle management tool designed to test autonomous AI agents at scale.
With synthetically generated data, teams can test Agentforce agents for accuracy, behavior, and actions while gaining full visibility into usage patterns and feedback.
🎥 Watch the demo to learn more!
Big advancements are coming to Flows!
Agentforce for Flow - Use AI to automate your process faster
Ever needed to create a flow in Salesforce but didn’t know where to begin? Or maybe you've faced the daunting task of auditing a massive, complex flow?
Agentforce for Flows is here to help you jumpstart the process with AI assistance:
-
Draft flows using natural language – Just describe what you want, and AI will build the foundation.
-
Generate formulas effortlessly – Tell it what to calculate, and it’ll do the rest.
-
Summarize complex flows – Get clear, human-readable explanations of existing flows.
Summary:
-
Accelerates, not replaces, the flow-building process.
-
Encourages adherence to best practices often skipped by human builders.
-
Supports an iterative, collaborative approach.
👉 Agentforce for Flows helps you go from stuck to started—fast. Check out the screenshots below to see it in action!
Summarize the existing Flow
Let AI build the first draft for you
Flow as an Agent Action
You just learned how AI can help you build and summarize flows, but wait a minute…. Flow IS THE BACKBONE of Agentforce, enabling scalable automation and seamless agent-human collaboration.
What is flow’s role in Agentforce?
- Automates repetitive tasks
- Standardizes agent workflow
- Reduces Agent build time
- Enhances end user experience
Now you can use Flow in Agentforce as:
- Agent’s Action
- Flow-invoked Agent
View this short demo on how to invoke an Agent from the Flow, and if you need more Flow and Agentforce Learning Resources, head on over to this link.
AI can look at your Org and draw process diagrams
If you’ve ever inherited a Salesforce org and had to audit or understand its configuration, you know the pain—it can take days, weeks, or even months, depending on the size of the org.
Now, you can do it in minutes.
Earlier in this post, we mentioned Elements.cloud—they’ve introduced a game-changing Process Configuration Mining tool that lets you visualize and track your entire Salesforce org’s configuration.
Honestly, it’s mind-blowing. My words won’t do it justice—just watch the presentation below.
Other Mentions You Shouldn't Miss
Can Agentforce Exist Without Data Cloud?
Technically, yes—you don’t need Salesforce Data Cloud to use Agentforce. But if you want your operations to run efficiently and intelligently at scale, integrating the two is the smart move. Here’s why:
-
🔗 Data Cloud as a Foundation: It acts as a centralized hub for customer data, giving Agentforce a unified view and enabling real-time identity resolution and enrichment.
-
🧠 Enhanced AI Capabilities: Agentforce taps into Data Cloud to leverage features like the Einstein Trust Layer for secure processing and RAG (Retrieval Augmented Generation) for contextual, relevant AI responses.
-
📊 Data Access & Context: Access to enterprise knowledge, customer data, and external sources means Agentforce can generate more intelligent, actionable insights.
-
🤖 Autonomous Actions: With data at its fingertips, Agentforce can automate tasks, trigger workflows, and personalize experiences.
But There’s More to Consider
If you're planning to take full advantage of Agentforce, Data Cloud integration comes with strategic considerations:
-
💸 Cost: It’s consumption-based, so plan your credit usage carefully.
-
👥 People: You’ll need data architects to design and manage structured and unstructured data pipelines.
-
📚 RAG & Unstructured Data: RAG is a game-changer—it allows Agentforce to retrieve and ground responses in large volumes of unstructured data (think PDFs, docs, past interactions). This means agents don’t just answer—they understand context and reference actual knowledge.
-
🛡️ Governance: Strong data governance is a must—ensure data is clean, compliant, and well-structured for reliable outcomes.
Prompt Builder
Behind every generative AI result is a prompt—a set of instructions.
Prompt Builder lets you create reusable prompt templates that can be used across your Salesforce org, including in Flows and AI Agents, to trigger actions and complete tasks.
Common Use Cases:
-
Summarize records
-
Generate content (e.g., write sales emails, create field summaries)
-
Enhance automation in Flows
To get started with Prompt Builder, check out the short demo. For in-depth training and resources, visit the Salesforce Prompt Builder center.
Agentforce Exchange
TDX introduced Salesforce's trusted marketplace and community for Agentforce, a suite of AI agents, where partners, developers, and the Agentblazer community can build, discover, and monetize AI components, enabling businesses to rapidly build and deploy AI agents.
Agentforce Developer Edition
Great news for developers: There's now a new Developer Edition org with Agentforce and Data Cloud enabled—no more juggling multiple short-lived orgs just to test your solutions. Use QR code below to sign up!
Agentforce API
You can now integrate headless agents in third-party apps thanks to the new Agentforce API. See the API capabilities through a practical demonstration.
No-Code LWC Builder Announced at TDX
Avonni introduced a game-changing No-Code LWC Builder—a visual, drag-and-drop tool that covers up to 90% of business needs. Easily nest components, build dynamic UIs, and see everything come to life on a visual canvas. It’s fast, flexible, and requires zero coding.
Check this live demo to have your mind blown.
Learn It. Hack It. Certify It. — Hands-On at TDX
At TDX, I took part in Hands-on Training, Mini-Hacks, and focused Bootcamps. These interactive sessions provided real-world practice, expert guidance, and opportunities to collaborate on live challenges. It was a fast track to learning, networking, and staying ahead with the latest Salesforce innovations.
Agentforce Hackathon 🤖 - my favorite part :)
Kicking off TDX with the Agentforce Virtual Hackathon was the best decision.
Not only was it my favorite part of the conference, but it also set the tone for everything that followed.
The Agentforce Virtual Hackathon brought builders together to create and showcase AI agents using the Agentforce platform—competing for prizes, recognition, and solving real-world business challenges.
But it wasn't just about hacking—it’s about the community, the creative energy, and yes, being entertained by the Salesforce Robotics Lab 🤖
Check out some highlights from the event below! 🎥📸
Final Thoughts
To get real value from Agentforce, you need a solid foundation in core Salesforce capabilities—like Sales and Service apps, Flows, Prompt Builder, and Data Cloud. These aren’t optional anymore—they’re essential.
There's still a noticeable gap between Salesforce’s ambitious AI vision and the readiness of many devs and admins to adopt it.
As one developer put it:
Only after addressing foundational issues should Salesforce start talking about Agentforce.” – Vuillamy
It’s fair to ask:
Is it too much too soon?
Will organizations have the budget for training, licenses, and the right talent to implement this at scale?