Conversational Fraud Investigation with SAS VI Network

Fraud investigators often work with highly complex financial networks containing thousands of entities and links. Manually exploring these networks to uncover relationships or anomalies or hidden patterns is slow and technically demanding. Traditional systems require investigators to know specific tools and workflows, creating dependency on technical teams and delaying case outcomes.

The SAS VI Network combines SAS Visual Investigator with a Large Language Model (LLM) agent and MCP (Model Context Protocol) framework. Through MCP, the AI agent intelligently selects and orchestrates the right SAS analytical tools — such as entity link analysis, transaction pattern detection, or network scoring — based on the user’s query. Investigators simply ask questions in natural language (e.g., “Show circular transactions linked to Taxpayer X”), and the system automatically identifies relevant tools, executes the analysis, and visualizes the results. This seamless AI–MCP synergy eliminates manual configuration, enabling effortless, context-aware fraud exploration.

• Smarter Tool Orchestration: MCP automates tool selection and execution, reducing reliance on technical expertise.
• Faster Investigation Cycles: Query-to-insight turnaround time dropped from hours to minutes.
• Higher Accuracy: Context-aware responses improve precision in identifying hidden patterns.