Enterprise Agentization: SAS and Anthropic Bring Specialized Agentic Workflows to Marketing and SMBs
Enterprise Agentization: SAS and Anthropic Bring Specialized Agentic Workflows to Marketing and SMBs
Summary
The “Agentic AI” era is rapidly maturing from simple chatbots to specialized, multi-step workflow automation. In mid-May 2026, two major industry moves underscored this trend: SAS expanded its Customer Intelligence 360 platform with a multi-agent system for marketers, while Anthropic launched “Claude for Small Business,” featuring 15 pre-built agentic workflows. These developments signal a shift toward “Data Intelligence,” where AI doesn’t just suggest content but autonomously orchestrates audience segmentation, campaign design, and financial reconciliation within existing enterprise and SMB software stacks.
What happened
At the SAS Innovate event, SAS announced a significant upgrade to SAS Customer Intelligence 360 (CI 360), transitioning it into a multi-agent ecosystem. The new architecture features a supervisory SAS 360 Agent that coordinates specialized agents for Journeys, Audiences, Search, and Content. Simultaneously, Anthropic launched Claude for Small Business, a package of native connectors (QuickBooks, HubSpot, etc.) and ready-to-run workflows. These workflows enable Claude to perform complex tasks like reconciling transactions or running end-to-end sales campaigns with minimal human intervention, moving beyond mere text generation to active process management.
Why it matters
This wave of “Enterprise Agentization” addresses the “execution gap” in digital transformation:
- From Assistant to Orchestrator: AI is now moving into the “doing” phase. Instead of a marketer manually setting up touchpoints, the SAS Journeys Agent builds production-ready code for the entire campaign.
- Vertical Specialization: By providing 15 ready-to-run workflows, Anthropic is lowering the barrier for SMBs who lack the technical resources to build custom AI agents.
- Bridging Data Silos: Both platforms emphasize deep integration with existing tools (Microsoft Teams for SAS, QuickBooks/HubSpot for Anthropic), allowing agents to act as the “connective tissue” between fragmented business data.
Evidence
- SAS Multi-Agent System: Includes specialized agents that can generate SAS code for journey execution and interact with federated data for audience creation.
- Anthropic’s 15 Workflows: Includes specific automations for closing monthly books, chasing unpaid invoices, and managing vendor documents.
- Interoperability: SAS announced deep integration with Microsoft Copilot, while Anthropic released native connectors for the most popular SMB software (Canva, DocuSign, Intuit).
- Human-in-the-Loop Governance: Both systems require explicit human approval before agents execute external actions (e.g., sending an email or processing a payment), ensuring safety and compliance.
Analysis
The shift from monolithic LLMs to multi-agent systems (MAS) reflects a deeper understanding of how work actually happens in an enterprise. A single model often struggles with the context-switching required to manage a marketing campaign from brief to deployment. By breaking these tasks into specialized agents (the “worker” agents) under a supervisory “manager” agent, SAS and Anthropic are mimicking human organizational structures. This “Agentization” is not just about automation; it’s about “Data Intelligence”—the ability to reason over proprietary data and execute actions in the systems where that data lives.
Practical takeaway
- For Marketers: Evaluate your current tech stack for “Agent-Readiness.” Platforms like SAS CI 360 are moving toward conversational campaign building; start documenting your workflows so they can be easily “translated” into agent prompts.
- For SMB Owners: Explore the Claude for Small Business workflows. Focus on high-friction administrative tasks like invoice reconciliation where agentic AI can provide immediate time savings.
- For IT Leaders: Prioritize governance and observability. As agents start interacting with core financial and customer systems, robust “human-on-the-loop” checkpoints become mission-critical.
Open questions
- Integration Complexity: How easily will these “ready-to-run” workflows adapt to highly customized or non-standard SMB data structures?
- Liability and Error Handling: Who is responsible if an autonomous agent misinterprets a marketing brief and deploys an incorrect campaign to thousands of customers?
- The Cost of Agency: Will the increased token consumption and API costs of multi-agent systems outweigh the efficiency gains for smaller businesses?