Explore how Salesforce boosted natural-language-to-Flow precision by swapping fine-tuned models for a restricted, multi-level DSL framework and much more.

The Salesforce Engineering Blog offers a deep dive into Salesforce technologies, providing technical insights, best practices, and real-world examples. Developers can explore topics such as Salesforce development, architecture, and integrations, gaining  knowledge to build innovative solutions on the Salesforce platform. With contributions from Salesforce engineers and industry experts, the blog serves as a resource for the Salesforce community.

Salesforce Engineering

Salesforce's Agentforce team replaced fine-tuned models with a constrained, multi-stage Domain-Specific Language (DSL) architecture to generate production-ready Flow automation from natural language. The new approach separates planning (Architect phase) from implementation (Developer phase), enforces strict validation at each stage, and auto-generates DSL rules from Flow metadata definitions. This eliminates hallucinations, handles 63+ Flow types including complex Screen Flows and custom actions, and scales to 461 billion monthly executions while improving accuracy and reducing deployment cycles from months to continuous updates.

How Agentforce Achieved Accurate Flow Generation Across 461B Executions