Build With Confidence: Inside the New Agentforce Builder
Summary
The new Agentforce Builder overhaul introduces a graph-based engine and a human-friendly scripting language called Agent Script, giving admins more precise control over AI agent behavior. It enables predictable, deterministic flows with clear logic by combining deterministic instructions and LLM prompts, making agents smarter and more reliable. Admins can build agents with two flexible views, Canvas and Script, and leverage an integrated AI Assistant for quick iterations and troubleshooting. Additionally, the builder offers preview features to simulate and understand agent decisions before deployment, significantly improving trust and transparency in agent development.
Takeaways
- Use Agent Script to define deterministic and context-aware agent behavior.
- Build and edit agents in Canvas view for ease or Script view for precision.
- Leverage the AI Assistant to get suggestions and debug scripts efficiently.
- Preview agent reasoning with mock data to validate behavior before deployment.
- Integrate agents seamlessly with Salesforce tools like Flow, Apex, and Prompt Builder.
Salesforce has launched the new Agentforce Builder, a brand-new experience for building agents powered by an enhanced version of the Atlas Reasoning Engine. For admins, this new builder represents a meaningful shift in how much control, clarity, and confidence you have when designing agent behavior. Previously, Atlas had a focus on retrieving and interpreting prompts. Based on the nature of the prompt, it would then determine the flow of custom actions in order to execute a solution. The new builder uses a graph-based engine, which means it’s able to have that flow defined, controlled, and previewed. The flow portions and conversational portions can be customized distinct from each other, allowing for more predictable and deterministic outcomes. This is also known as hybrid reasoning. While the fundamentals of Agentforce haven’t changed, how admins build and control agents has.