Mark Beyer

This session details how Agentic AI will impact existing data management architecture and technologies, which new use cases it enables in data management and engineering, and which skills will be needed or become obsolete. We’ll also cover how to prepare budgets, teams, and operating models for these changes. These are now valid, frequently debated questions as Agentic AI evolves.

One key to AI confidence is ensuring data as utilized by models is reasonable, representative and responsible. AI Ready data is a continuous effort to assure the data is appropriate--whether prepared prior to the AI model or as within the AI implementation itself. We discuss how active metadata is how data talks to models, models are validated by humans, and how humans communicate with models.

This session covers the use and output of Gartner’s AI-Ready Data Toolkit, which includes practices for both structured and unstructured data. The process develops metrics that “stack” as you progress from POCs to multicontext data use, operationalization and production support. The session also explains how to customize the toolkit with your own thresholds and readiness analysis.

As AI shapes business decisions, making unstructured data AI-ready is a key governance priority. The quality, accessibility and security of unstructured data directly determine the performance of AI applications, particularly for GenAI. To unlock its value for AI initiatives, data and business leaders should evolve their governance strategies to effectively manage, protect and utilize unstructured data, ensuring it is AI-ready while meeting compliance and security requirements.

Most metadata in 2025 will remain passive in approaches with stats, reports, schema and business-developed glossary terms. However, most organizations must grow their maturity in metadata management. We start with traditional metadata techniques — passive. With AI undergoing confidence issues and the demand to reduce risk, grow AI confidence and provide data assurance, active metadata becomes key.

In this session, two analysts will have a dynamic dialogue presenting how federation and self-service are reshaping data management. Together, they’ll debate the risks and rewards of centralization and decentralization, exploring how, through federation and self-service, organizations can balance control with empowerment, and offer an actionable strategy for navigating the evolving landscape of data management.

D&A leaders struggle to prioritize and justify data management spend, especially amid cloud-driven cost unpredictability. Value stream analysis links data production to direct and indirect business outcomes, driving quantifiable benefits. This session will link core research on cost management and FinOps with a means of using active metadata to measure value resulting in holistic cost optimization.

One key to AI confidence is ensuring data as utilized by models is reasonable, representative and responsible. AI Ready data is a continuous effort to assure the data is appropriate--whether prepared prior to the AI model or as within the AI implementation itself. We discuss how active metadata is how data talks to models, models are validated by humans, and how humans communicate with models.

This session covers the use and output of Gartner’s AI-Ready Data Toolkit, which includes practices for both structured and unstructured data. The process develops metrics that “stack” as you progress from POCs to multi-context data use, operationalization and production support. The session also explains how to customize the toolkit with your own thresholds and readiness analysis.

This session covers the use and output of Gartner’s AI-Ready Data Toolkit, which includes practices for both structured and unstructured data. The process develops metrics that “stack” as you progress from POCs to multicontext data use, operationalization and production support. The session also explains how to customize the toolkit with your own thresholds and readiness analysis.