CData targets development success with Connect AI update
The vendor's latest launch takes a three-pronged approach toward helping users move pilots into production and could be a competitive differentiator if it delivers on its promise.
With most AI initiatives never reaching production, CData on Monday introduced updates to its Connect AI platform that add greater connectivity, context and control to help customers more successfully develop AI tools such as agents and chatbots.
CData unveiled the improvements, which are generally available, at the Gartner Data & Analytics Summit, a data and analytics industry conference hosted by the research and advisory firm in Orlando.
First launched in September 2025, Connect AI is a fully managed platform based on the Model Context Protocol open standard that is designed to simplify integrating AI models, including large language models such as OpenAI's ChatGPT, with an enterprise's proprietary data to inform AI applications.
Improved connectivity now provides read-write access to more than 350 data sources without replication or data movement, upgraded context includes three new tool types to enable agents and other AI applications to interact with information sources to carry out specific tasks, and enhanced control includes new security and governance specifications.
Collectively, the Connect AI update is significant for CData customers given that it addresses some of the problems data and AI teams face when trying to build AI tools, according to Stephen Catanzano, an analyst at Omdia, a division of Informa TechTarget.
"CData's update is very big for users as it addresses critical bottlenecks in AI production by integrating connectivity, context, and control into a single platform, enabling seamless access to live data, enhanced semantic intelligence and robust governance for production-grade AI deployments," he said. "Data connectivity from sources to AI is critical and [CData has] 350 connectors to assist."
Based in Chapel Hill, N.C., CData provides a data integration platform that includes connectivity to hundreds of data sources. Competitors include fellow specialists such as Fivetran and Informatica.
CData's update is very big for users as it addresses critical bottlenecks in AI production by integrating connectivity, context, and control into a single platform, enabling seamless access to live data, enhanced semantic intelligence and robust governance for production-grade AI deployments.
Stephen CatanzanoAnalyst, Omdia
However, despite enterprises investing more each year in building AI tools, the overwhelming majority of AI initiatives never make it past the pilot stage and into production. While lack of proper alignment between a project and its goal and outdated technology are among the reasons many AI projects fail, poor data quality and lack of relevant data are also prominent prohibitors.
In response, numerous data management and analytics vendors -- including Databricks, Domo and MongoDB -- have recently introduced capabilities specifically designed to help their users more successfully discover and operationalize high-quality, relevant data for their AI projects.
Now, CData is doing the same with its new Connect AI capabilities, the development of which was inspired by observing organizations consistently fail to move AI projects into production, according to Ken Yagen, the vendor's chief product officer.
"The model works, the proof-of-concept impresses everyone, and then it stalls because the infrastructure between the agent and the enterprise data isn't production-grade," he said. "The gap [is] the data access and orchestration layer. Agents need to reach live systems behind the firewall, understand the data they're looking at in business terms, and operate within guardrails."
CData's approach to helping customers more successfully move AI projects from pilots to production using its AI Connect platform is three-pronged:
Connectivity now includes read-write access -- the ability to view and modify data -- to more than 350 external systems without requiring users to move data, and On-Premise Agent to extend access to data regardless of where it resides.
To improve the context of AI pipelines, which already featured a semantic layer, CData is adding new tool calling capabilities including Universal Tools to work across all of an enterprise's data sources, Source Tools to more precisely control the requests an agent can make on data sources, and Custom Tools to enable access to specific data for explicit use cases.
New control measures include new authentications and permissions.
Perhaps the most valuable of the new Connect AI capabilities are the controls given that many enterprises have already figured out to connect data and add context, according to Michael Ni, an analyst at Constellation Research. Collectively, the Connect AI update is significant because it provides capabilities that help CData customers securely deploy agents that leverage proprietary data, Ni continued.
"The models aren't the problem," he said. "The problem is the data access, context and governance. Platforms that can provide secure, context-aware access to live enterprise data will determine whether AI pilots evolve into trusted operational systems. CData is introducing structural mechanisms for operationalizing AI agents safely against enterprise data, not just connecting them."
Beyond being beneficial for CData users, the Connect AI update helps distinguish the vendor from other data integration platforms by providing a connective layer between agents and enterprise data, according to Ni.
"Vendors that control the context layer between agents and business applications will shape the next generation of enterprise automation," he said. "CData [is] positioning itself as the connective layer between enterprise systems and AI agents, bridging a gap that traditional data pipelines and iPaaS platforms weren't designed to fill."
Catanzano cited the contextual improvements as the most valuable for users, noting that they directly address the accuracy of AI outputs and reduce unnecessary tool calls from agents and other AI applications that can lead to errors.
Meanwhile, from a competitive standpoint, he added that AI Connect could help CData differentiate itself from other vendors if it delivers its intended high accuracy and security.
"CData's Connect AI update differentiates it by combining high accuracy with production-grade security and semantic intelligence, positioning it ahead of competitors who may still rely on simpler API-based approaches," he said. "This is the real value-add to move AI projects from pilots to production."
Next steps
Beyond the new capabilities introduced on Monday, CData's roadmap includes continuing to expand connectivity to more data sources, developing capabilities that enable agents to execute prebuilt multi-step workflows for common use cases and improving the vendor's semantic layer, according to Yagen.
"Right now, Connect AI provides prebuilt semantic context for popular SaaS applications so agents understand the data they're working with. Future releases will expand that to give AI systems a richer, business-aware understanding of enterprise data structures," he said. "The goal is for agents to reason about data, not just access it."
Focusing on adding greater context is wise, according to Ni, who noted that rich semantic modeling could improve CData's standing relative to its competition.
"The opportunity is to own the connective tissue between enterprise systems and the AI context layer," he said. "That means exposing richer operational metadata, extending high-performance operational queries and enabling policy-driven agent actions so AI systems can interact with enterprise systems reliably and safely."
Catanzano, meanwhile, suggested that CData could better serve existing customers and perhaps attract new ones by adding scale and breadth to its offerings to improve support for AI initiatives.
"CData could focus on expanding its platform's scalability, adding support for emerging AI use cases, and deepen partnerships with enterprise software providers to further enhance integration capabilities," he said.
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than three decades of experience. He covers analytics and data management.
