SAS announced major expansions to its Viya platform at the SAS Innovate conference, coinciding with the company’s 50th anniversary. Three new capabilities move the enterprise analytics platform into agentic territory: a Model Context Protocol (MCP) server that exposes SAS analytics as tools for external AI agents, embedded AI copilots across the analytics lifecycle, and a framework for building and deploying governed agents at scale.

MCP Server: SAS as a Callable Tool

The SAS Viya MCP Server uses the open MCP standard to expose SAS analytics, models, and decisioning capabilities to external AI agents. According to SAS, organizations can embed SAS analytics, governance, and domain intelligence into their own agents from the LLM interface of their choice, including Claude and ChatGPT, “without duplicating logic or bypassing enterprise controls.”

This is the pattern gaining traction across enterprise software: instead of rebuilding analytics capabilities inside LLMs, expose existing production systems as agent-callable tools via MCP. The agent handles reasoning and orchestration; SAS handles the analytical execution within its existing governance boundary.

Embedded Copilots, Not Chatbot Overlays

SAS Viya Copilot is a family of AI assistants embedded directly into production analytics workflows, integrated with Microsoft Foundry. As Techzine reports, the copilot “truly operates from within the analytics environment itself and does not function as a separate chat interface alongside it.”

Current capabilities span general Q&A across core Viya applications, AI-generated SAS and Python code with documentation, model pipeline guidance, conversational dashboarding, and visual investigation with AI-powered case narratives. Two industry-specific copilots are already available: one for Asset and Liability Management (ALM) in financial risk workflows, and one for Health Clinical Data Discovery in clinical data analysis. SAS plans to expand to financial crime prevention and supply chain optimization later in 2026, according to its press release.

Agentic AI Accelerator

The SAS Agentic AI Accelerator provides a curated framework of code, components, interfaces, and best practices spanning no-code, low-code, and full developer workflows. It targets teams building and deploying AI agents within SAS Viya, aiming to move organizations from experimentation to governed production deployment.

Alongside the accelerator, SAS also announced the SAS Retrieval Agent Manager (RAM), a no-code RAG solution for transforming unstructured data into context-aware AI responses. SAS plans to integrate RAM into Viya to ground agent and assistant outputs in enterprise context, per the company’s announcement.

The Governance-First Bet

“The role of human expertise in operationalizing agentic AI is not diminished by automation; it’s elevated,” said Jared Peterson, Senior Vice President of Global Engineering at SAS, in the press release.

SAS is betting that domain-specific, governance-first agent infrastructure will outperform generic LLM assistants in analytics workflows where regulatory compliance, data lineage, and auditability are non-negotiable. The MCP server approach lets SAS participate in the broader agent ecosystem without ceding control of its analytics execution environment to third-party LLMs. For the growing list of enterprise vendors adopting MCP as their agent interoperability layer, SAS is further validation that the standard is becoming the default bridge between legacy enterprise systems and the agentic stack.