Overview of AI-Powered Features in TimeXtender

TimeXtender's AI-powered features include the MCP Server, Xpilot Data Quality, and Xpilot Orchestration Error Insights. This document explains where data is processed and hosted to help you make informed decisions about deployment in your environment.

All AI features in TimeXtender are optional. You have complete control over which features to enable and how they're configured.

Our Philosophy: Customer Choice

TimeXtender's approach to AI features is built on the principle of customer choice and control:

Optional by Design

All AI features are optional. You decide which features are enabled, when to enable them, and how they're configured. TimeXtender's core data integration capabilities work completely independently of these AI features.

Flexibility in Deployment

We provide multiple deployment options to meet your specific security, compliance, and operational requirements. You can choose between cloud-based AI services or fully on-premises solutions depending on your needs.

Transparency

We believe you should know exactly where your data is processed. This document provides complete transparency about data flows, processing locations, and available options for each feature.

Your Data, Your Control

• Your data is never used to train AI models
• You maintain complete control over what data is processed
• Features can be disabled at any time without affecting your existing workflows

TimeXtender MCP Server

The MCP Server can run fully locally if needed:

• The MCP Server component runs as a Windows service in your environment
• Your data remains in your Prepare instances
• Language model options:
  • Cloud-based: Deploy cloud-based agents like Claude for fast and scalable operation
  • Fully local: Use local models (e.g., LM Studio) on your own hardware - no external data transmission

What this means:

• If you run a local language model, all processing stays on-premises with no data leaving your environment
• If you use cloud-based models, only semantic metadata (table/field names, descriptions, relationships) and query results are sent to the AI service for processing
• The MCP Server enforces read-only access regardless of configuration

MCP Server Architecture: Zero-Access Security

The TimeXtender MCP Server is built on the same zero-access security model that governs the entire TimeXtender platform: it orchestrates using metadata and never accesses, moves, or stores your actual data.

The architecture enforces strict security boundaries at every layer:

• Prepare Instance Storage (Your Warehouse): Your data stays exactly where it is — in your own Snowflake, Azure SQL, Fabric, or AWS environment. The MCP Server connects through read-only database authentication, ensuring it can query but never modify your data.
• Semantic Model (MCP Server): The MCP Server runs as a Windows service inside your environment. It translates AI requests into safe, read-only queries using your semantic layer — table names, field descriptions, and relationships. No raw data is stored or cached by the server itself.
• Self-Hosted AI Tools & Agents: AI clients (Claude, ChatGPT, LM Studio, or custom agents) connect to the MCP Server via API key authentication. You control which AI tools have access and can revoke credentials at any time.

Two distinct security boundaries - database authentication on one side, API key authentication on the other - with the MCP Server acting as a governed intermediary. Your data never leaves your environment. The AI only sees what the semantic layer exposes, and only in read-only mode.

Xpilot Data Quality (Release 26.1)

Currently processes through OpenAI:

• Schema information, field descriptions, and sample data (first 100 rows) are sent to OpenAI for analysis
• AI generates suggested validation rules as drafts
• You review and approve rules before they're published

Next release: Will process through Microsoft Azure OpenAI services in Europe

Xpilot Orchestration Error Insights (Release 26.1)

Currently processes through OpenAI:

• Error logs and execution context are sent to OpenAI for analysis
• AI provides diagnostic summaries and resolution suggestions
• No business data is included in error analysis

Next release: Will process through Microsoft Azure OpenAI services in Europe

Future Roadmap: Enhanced Regional Control

Next Release (Planned)

Microsoft Azure OpenAI Integration (EU Region)

• Xpilot Data Quality and Xpilot Orchestration Error Insights will process through Microsoft Azure OpenAI services hosted in European data centers
• No change to functionality or user experience

Future Releases

Regional Alignment

• AI processing regions will align with your TimeXtender TDI instance region
• Reduces latency and simplifies compliance