Financial decision intelligence platform MindBridge has introduced MindBridge MCP, a new capability that connects the platform’s AI-powered financial intelligence directly to supported AI assistants, including Claude, Microsoft Copilot, and Cursor.
MCP makes it easier for finance and audit professionals to access and investigate financial risk without leaving the AI environments where they’re already working, MindBridge says.
Through a secure connection to a customer’s MindBridge environment, authorized users can ask questions in natural language and retrieve insights from completed MindBridge analyses directly within their AI assistant.
For example, an auditor could ask: “Show me high-risk transactions over $100,000 from my 2025 year-end general ledger analysis and explain which control points contributed to their risk scores.”
Rather than manually navigating tables and exporting results, the user can work through the investigation conversationally, using the analysis already performed by MindBridge. This brings together two complementary capabilities: MindBridge’s independent, explainable financial risk analysis and the conversational experience of modern AI assistants, the company says.
The result is a more accessible way to investigate financial activity and understand potential risks.
The initial release supports four core capabilities:
- Finding relevant analyses: Discover completed MindBridge analyses available to the user, including searching by keyword.
- Understanding risk results: Explore the structure of an analysis, including available fields, risk scores, and control points.
- Investigating why transactions were flagged: Access information about the control points contributing to transaction risk scores, helping users interpret findings.
- The query of financial data: Retrieve specific results, such as high-risk transactions above a defined monetary threshold, without manually exporting data.
Each user connects individually through authentication and consent, and access remains governed by their existing MindBridge permissions. An AI assistant can only retrieve information that the authenticated user is authorized to access, the company says.
The initial release is also read-only. Users can find analyses, explore risk information, and query results, but can’t create, modify, or take action on MindBridge records through MCP.
Organizations retain control over adoption. MCP is disabled by default and must be explicitly enabled by a tenant administrator, who must also acknowledge a risk disclosure before activation, MindBridge says.
MindBridge MCP is now generally available to customers as part of the 26.10 release. More information can be found here.
Photo credit: MindBridge/LinkedIn
Sign in to get access to this free resource, and all of our whitepapers and reports.
Download this content today!
Register Now Already registered? Click here to Log In
