Hello SAS Community,
Many of us are exploring how to use Large Language Models (LLMs) for more than just simple chat. The logical next step is building systems that can perform multi-step tasks - often called AI agents - systems that go beyond basic chat and can perform multi-step tasks, orchestrate workflows, and drive decisions. However, integrating these complex systems into existing enterprise environments — while ensuring governance, scalability, and security — can be a real challenge.
That’s exactly why we built the SAS Agentic AI Accelerator.
🎥 Watch our demo video to see it in action
🧠 What Is the SAS Agentic AI Accelerator?
The SAS Agentic AI Accelerator is a curated collection of code, components, interfaces, and best practices to help you build, govern, and deploy AI agents within your existing SAS Viya environment - with no unsupported hacks or workarounds.
It enables organizations to go from a promising AI concept to a production-ready solution - in a structured, repeatable, and governable way.
👥 Designed for Every Role: No-Code, Low-Code, and Yes-Code
One of the most powerful aspects of the accelerator is its multi-interface design - tailored for different skill sets and teams:
- For Analysts (No-/Low-Code):
Use a clean, no-code UI for prompt engineering or leverage SAS Studio Custom Steps to build agents with guided low-code tools.
- For Developers (Yes-Code):
Access SAS Macros for advanced control and explore a ready-made Postman collection to test agent APIs and integrations.
- For Decision Managers:
Use the custom node for SAS Intelligent Decisioning to embed agent logic directly into your decision flows.
🧰 What’s in the Package?
Here’s what you’ll get in the Agentic AI Accelerator:
✅ Deployment-Ready Code and Docs
Step-by-step instructions and a complete codebase to get you up and running in your Viya environment.
✅ Pre-Built Interfaces
Flexible tools for analysts, developers, and decision managers — all within a secure, auditable framework.
✅ Model Deployment Recipes
Quick-start templates for deploying open-source LLMs and embedding models, enabling capabilities like Retrieval-Augmented Generation (RAG).
✅ Built for the Enterprise
No experimental APIs. No shortcuts. The accelerator is built entirely on standard SAS Viya capabilities, ensuring long-term support with future Viya updates and releases, integration with existing security policies, and compatibility with your IT stack.
💡 Ready to Build Your First Agent?
You can download the full accelerator package now and try it out in your own test environment:
👉 Get started here:
https://github.com/sassoftware/sas-agentic-ai-accelerator
We’d love to hear how you’re using it - whether you're experimenting with a new agent use case or embedding it into an existing decision flow. Share your questions, ideas, and feedback in the SAS Community or directly through the GitHub repository.
Let’s build the future of intelligent automation together.
