SAS has been named a Major Player in the IDC MarketScape: Worldwide Agentic AI Platforms 2026 Vendor Assessment, a recognition that puts its analytics and decision-intelligence capabilities in the spotlight as enterprises evaluate platforms for building and governing AI agents. The company’s SAS Viya platform combines agent development tools, analytics, decisioning, model governance and integration capabilities. Its approach emphasizes combining large language model (LLM) reasoning with deterministic business rules, aiming to make agent-driven decisions more explainable and controllable in complex enterprise environments.
SAS Builds Its Agentic AI Strategy Around Decision Intelligence
The recognition reflects SAS’s attempt to extend its established analytics and machine learning business into the emerging agentic AI platform market. Rather than positioning Viya solely as a framework for creating autonomous agents, the company is bringing together data management, predictive analytics, optimization, business rules and AI governance.
The platform combines four principal components: SAS Viya Agent Builder for creating agents, SAS Viya Copilot for AI-assisted interaction, an open-source Model Context Protocol (MCP) server exposing more than 40 analytical tools, and SAS AI Navigator for governance and oversight. IDC describes this combination as agentic capabilities layered onto a cloud-native analytics and decisioning platform.
The architecture addresses a challenge facing enterprises as AI systems move from generating recommendations to executing tasks. An agent may need to retrieve customer information, evaluate risk, apply business policies and initiate an action. Each step introduces potential errors or compliance concerns if the system relies entirely on an LLM’s probabilistic output.
Hybrid AI Connects LLMs With Deterministic Rules
A central element of SAS’s approach is what the company calls hybrid AI. It combines LLM-based reasoning with deterministic business rules, optimization techniques and validated analytical models.
The distinction is particularly relevant in regulated workflows. A language model can interpret a request or help an agent determine which information it needs. A deterministic rule or validated model can then apply predefined thresholds, eligibility conditions or risk calculations.
For example, in banking, an AI agent could help assemble information for a lending decision while established credit policies and validated risk models govern the actual assessment. In insurance, similar architecture could support claims processing while retaining explicit business rules and human review for exceptions.
This approach does not eliminate model errors or guarantee correct decisions. Its value depends on how well the underlying rules, data, models and approval processes are designed. However, it gives enterprises a way to constrain AI-assisted workflows using logic that can be documented and reviewed.
IDC identified decision intelligence as one of SAS’s principal strengths, alongside agent quality assurance and deployment flexibility.
Governance and MCP Expand the Platform’s Role
SAS is also emphasizing controls across the agent development lifecycle. IDC highlighted capabilities including explainability, model cards, model and context lineage, fairness monitoring, human-in-the-loop gates, lifecycle auditing and an inventory of AI systems through AI Navigator.
These capabilities matter because enterprise agents can interact with multiple applications and data sources. Organizations need to know which models and information an agent used, what actions it performed and where human intervention is required.
The MCP server adds an interoperability layer. MCP is an open protocol designed to connect AI applications and agents with external tools and data sources through a standardized interface. SAS says its server allows agents built using external frameworks, including Microsoft Copilot, Claude and LangChain, to access governed Viya analytics and decisioning services.
This could help organizations avoid making every agent dependent on a single development environment. Enterprises can use different models or agent frameworks while retaining access to selected analytics and decision services.
IDC also highlighted deployment flexibility. Its assessment lists managed cloud deployments on AWS, Microsoft Azure and Google Cloud, alongside private cloud and on-premises options. That flexibility is relevant for organizations with data residency, sovereignty or infrastructure requirements that make public-cloud-only architectures difficult.
Agentic AI Platforms Face a Growing Governance Test
SAS is entering a competitive market that includes enterprise software vendors, automation providers, hyperscalers, data platforms and specialist AI companies. Microsoft, Google, Amazon Web Services, Salesforce, IBM, ServiceNow, Dataiku and Camunda are among the vendors included in IDC’s assessment.
The market is developing quickly. IDC reports that, as of July 2026, 95% of organizations had at least one agent workflow in production, while nearly one-quarter had more than 15. Its research also found that transparency and explainability were challenges for 37% of organizations implementing agentic AI, while 36% cited implementation and integration costs.
Those figures help explain why enterprise buyers are evaluating more than agent-building tools. They need platforms that can manage identities, permissions, integrations, monitoring and operational performance after deployment.
Importantly, IDC’s assessment also identifies areas where SAS needs to improve. The report says agent evaluation and continuous improvement capabilities were still basic at the assessment cutoff, although enhancements were on the roadmap. It also describes support for agent identity, quarantine, advanced action and reasoning guardrails, multi-agent permissioning and advanced logging as only partial. SAS
These limitations qualify the recognition. SAS’s analytics and decisioning heritage may be attractive to regulated organizations, but buyers seeking mature runtime security, extensive agent observability or developer-first flexibility will need to assess the platform against their specific requirements.
IDC’s assessment also distinguishes capabilities available by March 1, 2026, from features that were still planned or unreleased. This matters when comparing current product functionality with longer-term platform strategy.
The Enterprise AI Opportunity Moves Beyond Agent Creation
For SAS, the Major Player designation reinforces its position in a market where trusted decision-making is becoming as important as model intelligence. Its combination of analytics, business rules, governance and deployment choice is aimed particularly at banking, insurance, healthcare, government and manufacturing.
The broader enterprise challenge is moving from isolated agents to coordinated systems that can perform business tasks under defined controls. As organizations expand agent deployments, platforms will be judged not only by how quickly they can build an agent, but by whether they can explain its decisions, restrict its authority and demonstrate that it operates within business and regulatory requirements.
SAS Viya is designed to address that combination. Its longer-term competitiveness will depend on how quickly the company closes the runtime security and evaluation gaps identified by IDC while making its decision-intelligence capabilities accessible across increasingly diverse agent ecosystems.
Market Landscape
The agentic AI platform market is shifting from agent creation toward lifecycle management, orchestration and enterprise governance.
- Enterprise platforms: Microsoft, Google, AWS, Salesforce and ServiceNow are extending their cloud and business application ecosystems into agent development and orchestration.
- Automation providers: IBM, Camunda, UiPath and Automation Anywhere are bringing process automation and integration capabilities into agentic workflows.
- Data and analytics platforms: SAS, Dataiku, DataRobot, Databricks and Snowflake are emphasizing data foundations, model evaluation and analytics integration.
- Governance and security: Agent identity, access permissions, auditability, observability and human oversight are becoming essential platform requirements.
- Hybrid deployment: Private cloud, on-premises and model-agnostic options remain important for regulated and data-sensitive industries.
IDC’s assessment reports that 62% of organizations consider sophisticated guardrails a key capability for agentic AI platforms. It also found that 42% are targeting private cloud stacks and 34% bare-metal servers for agent deployments. SAS
Top Insights
- SAS was classified as a Major Player in IDC’s 2026 assessment, which evaluates agentic AI platform capabilities and longer-term vendor strategies.
- SAS Viya combines Agent Builder, Viya Copilot, an MCP server with more than 40 analytical tools and AI Navigator.
- Hybrid AI combines LLM reasoning with deterministic rules, optimization and validated models to support auditable enterprise decisions.
- IDC identified agent evaluation, runtime security, identity and advanced guardrails as areas where SAS still has capability gaps.
- Deployment flexibility and analytics integration could make SAS relevant to regulated enterprises seeking governed agentic AI infrastructure.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
