CData Software, the data layer for AI, announced major enhancements to CData Connect AI at the Gartner Data & Analytics Summit. The updates extend CData’s managed Model Context Protocol (MCP) platform with new capabilities across connectivity, context, and control, the three pillars required to move AI from experimentation to production.
Why AI Stalls Before Production
AI investment is accelerating. “Gartner says worldwide AI spending will total $2.5 trillion in 2026.” But spending isn’t translating into results. Most generative AI initiatives still stall before reaching production. The bottleneck isn’t model capability, it’s the data infrastructure underneath. Without live connectivity to business systems, semantic intelligence that gives data context to AI, and governance controls that enforce security at scale, AI initiatives fail to deliver business value.
CData’s own State of AI Data Connectivity Report reinforces this reality. Only 6% of organizations are satisfied with their current data infrastructure for AI. More than half still rely on custom-built integrations that can’t scale. And 71% of AI teams spend over a quarter of their implementation time on data integration alone, time spent wiring plumbing instead of building intelligence.
“AI agents are only as effective as the tools they can access and the data behind them, and only as safe as the controls governing both,” said Amit Sharma, CEO and Founder of CData. “This release gives teams the ability to build use-case-specific agent tools with the right business context, deploy them securely, and enforce granular controls over which data agents can access, which actions they can take, and under what identity. That’s what’s been missing, not better models, but the connectivity, context, and control that make agents trustworthy enough to run in the enterprise.”
Also Read: Dfns Launches Payouts
Connect AI: Connectivity, Context, and Control in a Single Platform
CData Connect AI is purpose-built to address the data infrastructure gaps that prevent AI from reaching production. Today’s enhancements extend the platform across all three pillars:
Connectivity: Connect Gateway and 350+ Data Sources
Connect AI provides live, read-write access to more than 350 business systems, without replication or data movement. The new Connect Gateway extends this reach to data sources behind the firewall, with support for SAP, SQL Server, and PostgreSQL, and more. The result: AI systems can operate against live data regardless of where it resides.
Context: Expanded Agent Tooling and Toolkits
AI agents need business-aware context to choose the right actions and avoid unnecessary MCP tool calls. But exposing too much context creates new risks: increased token usage, model confusion, and unintended access to sensitive data or operations. Connect AI addresses this challenge with a scoped MCP architecture that precisely controls what each agent can see and do. This release introduces three complementary tool types:
- Universal Tools provide a normalized set of operations that work consistently across all 350+ connected systems. Instead of exposing hundreds of system-specific tools, agents receive a compact, schema-aware interface ideal for data exploration, ad-hoc analysis, and multi-source reasoning – without tool surface bloat.
- Source Tools expose tightly defined operations specific to each system. These tools map directly to approved system actions, allowing IT teams to enforce predictable execution, transactional safety, and auditability for production workflows.
- Custom Tools allow organizations to define purpose-built operations tailored to specific workflows. These tools execute pre-optimized queries with explicit data access limits — reducing token usage, improving performance, and eliminating unintended data exposure.
Workspaces define the data boundary for each agent by specifying exactly which datasets, schemas, or views are accessible. New Toolkits define the action boundary by determining which Universal, Source, or Custom Tools are available. Each Workspace and Toolkit combination can be deployed as a dedicated MCP server, ensuring that agents operate only within their intended scope; reducing context noise, strengthening governance, and delivering enterprise-grade control over agent behavior.
SOURCE: CData