As AI capabilities evolve, many teams struggle to balance the flexibility of Large Language Models (LLMs) with the reliability and predictability required for production systems. How do you harness generative AI without sacrificing control, transparency, or trust?

In this session, we’ll explore how Salesforce’s new graph-based Agent Script helps bridge that gap—combining the adaptability of LLMs with deterministic, structured workflows.

This is a hands-on, practical session focused on real-world use cases rather than theory. You’ll learn how graph-based agent design enables clearer logic, better observability, and safer execution paths, while still leveraging AI-driven reasoning where it adds the most value.

What you’ll learn:

How graph-based Agent Script differs from traditional linear workflows

When to use LLM reasoning vs. deterministic steps

How to design reliable, debuggable agent flows

Patterns for combining automation, decision logic, and AI responses

Whether you’re just getting started with AI in Salesforce or already experimenting with agents, this session will give you concrete ideas and patterns you can take back to your org.

Join us to learn how to build smarter, safer, and more reliable AI-powered workflows—without losing flexibility.