At MadConnect, we leverage Artificial Intelligence (AI) to enhance the efficiency, accuracy, and connectivity of our Intelligent Connectivity Layer (ICL). We are committed to transparency regarding how these technologies interact with your data and our platform.
Currently, MadConnect integrates with OpenAI to power specific automation features. In the future, we may expand our ecosystem to include other leading model providers (such as Anthropic or Google Gemini) to ensure the highest level of service.
Our AI integration is focused on two primary functional areas:
1. Intelligent Schema Mapping
To streamline the onboarding of complex data sets, our ICL application uses AI to:
Review Client Sample Data: Analyze the structure and headers of uploaded sample files.
Recommend Mappings: Automatically suggest how client-specific data fields should map to ICL standard schemas, reducing manual configuration time and potential errors.
2. Agentic Control Plane via MCP
We utilize the Model Context Protocol (MCP) to allow AI agents to interact securely with our application APIs. This enables:
Automated Configuration: Using agents to set up and configure new data connections based on natural language instructions.
Minimal Exposure: We only share the metadata or sample records necessary to perform the requested mapping or command. But this is under the user’s control to share data or use an MCP server.
No Training on Client Data: We utilize enterprise-grade API agreements to ensure that client data processed through these AI models is not used to train the providers’ foundational models.
User Oversight: AI recommendations for schema mapping and connection setups are presented as suggestions that can be reviewed and overridden by the user.
