Actian, the data and artificial intelligence division of HCLSoftware, has announced the launch of the Actian Data Steward Agent. Embedded directly within the Actian Data Intelligence Platform, this newly developed autonomous AI agent establishes a managed semantic layer to deliver a unified, shared business context across enterprise AI frameworks, internal workflows, Model Context Protocol (MCP) connected tools, and third-party AI agents.
By automating the labor-intensive tasks associated with metadata documentation, enrichment, and governance, the agent accelerates time-to-value for corporate data investments. It significantly reduces the manual developer effort traditionally required to construct and preserve AI-ready data foundations.
Also Read: Unstructured Deepens Microsoft Azure Integration to Streamline Enterprise AI Infrastructure
Data and analytics leaders frequently struggle to maximize the utility of their information assets. According to Gartner®, “ majority of data and analytics (D&A) leaders still fail to leverage metadata effectively, as 51% of organizations still rely on passive metadata practices, limiting their ability to unlock business value. ” Consequently, data cataloging and maintenance burdens typically fall squarely on human data stewards, creating operational bottlenecks. Gartner also notes that “ these stewards often have only a fraction of their time typically 5% to 10% dedicated to governance tasks, including cataloging. ”
As corporations rapidly integrate autonomous AI agents into daily operations, maintaining uniform metadata has become an essential prerequisite for system accuracy. The Data Steward Agent natively embeds into data catalog workflows to dynamically update metadata as organizational data landscapes evolve.
Architectural Design and Core Integration
The agent functions via a decentralized infrastructure that links directly with the distributed knowledge graph, semantic layer, data products, and data contracts of the Actian Data Intelligence Platform. This unified architecture ensures that both native platform features and external AI systems including those linked via Model Context Protocol (MCP) or agent-to-agent (A2A) protocols operate under an identical, governed business context.
By continuously identifying documentation gaps, recommending structural updates, and enforcing semantic uniformity, the tool eliminates tedious manual overhead. This allows data engineering teams to pivot from manual data entry toward high-value validation and scalable architecture curation.
End-to-End Metadata Automation
While conventional AI copilots are limited to executing isolated, reactive commands, the Data Steward Agent uses agentic AI to oversee the comprehensive lifecycle of a data steward’s portfolio. The software continuously tracks the state of the data infrastructure and automates key catalog processes:
- Dynamic Resource Documentation: Authoring and refreshing data asset descriptions in real time as data pipelines morph.
- Alignment of Asset Ownership: The system finds dormant data assets on its own and shows the missing pieces in corporate documentation or ownership chains.
- Smart Data Classification: Suggesting classification tags, structural ones even, for example PII sensitivity of risk levels, and operational domains.
- Updating and Maintaining the Corporate Glossary: Creating the standard, company-wide glossary of definitions and terminology; keeping it updated.
- Ahead-of-Time Compliance Monitoring: Scanning at a high level to identify instances where policy violation is likely before the wrong data formats are sent to applications downstream.
Governance-Ready Data Precision
Utilizing advanced natural language processing (NLP), the Data Steward Agent formulates its metadata recommendations by evaluating data lineage, operational usage patterns, and preexisting data catalog context to maximize output precision. Every generated recommendation aligns with existing data products and contracts enforced within the platform to maintain system compatibility. Human data stewards retain ultimate authority, shifting their operational role from content creation to high-level verification to preserve strict corporate oversight and internal controls.
“As enterprises deploy connected AI agents across their workflows, manually maintaining semantic consistency becomes operationally impossible,” said Guillaume Bodet, Chief Product Officer at Actian. “The Data Steward Agent continuously aligns metadata, business definitions, provenance, and governance context so that AI systems can operate on a shared understanding of enterprise data.”
