Details
This webinar will examine the pivotal role of data products in fueling the AI revolution. We’ll explore how AI not only powers the creation and consumption of data products, but also relies on robust metadata to ensure trust and reliability. The discussion will highlight the use of Generative AI to automate data product development, along with essential techniques for safeguarding design integrity. We’ll also cover how natural language interfaces are transforming the way organizations interact with data products, concluding with a case study presented by the Chief AI & Data Officer of digital-first, India-based financial innovator L&T.
- Finance, will be used to illustrate these points.
- Learn about quantification of Return-On-Data Investments (RODI).
- Strategies for future-proofing AI deployments across evolving cloud and LLM landscapes.
Learn how business users can be empowered to create data products through a scalable manufacturing process using Generative AI. - Discover how business users can be empowered for self-service access to data using natural language interfaces via Generative AI.
Speakers
Post-event summary
The webinar titled “AI Needs Data Products (and Data Products Need AI),” was hosted by EDM Association and Ab Initio and explored the deep interdependence between artificial intelligence and data management in modern enterprises. Featured speakers included:
- Stephen Brobst, Chief Technology Officer, Ab Initio Software
- Dr. Debarag Banerjee, Chief AI and Data Officer, L&T Finance
- Moderator: Jim Halcomb, Chief Research & Development Officer, EDM Association
The discussion centered on how effective data product architectures and governance frameworks underpin the success of AI initiatives, and conversely, how AI can now automate and enhance the creation and management of those same data products.
Stephen opened the session by emphasizing that “AI without the data is just artificial,” underscoring that intelligence depends on high-quality, well-structured, and trustworthy data. He explained that a true data product goes beyond raw information. It must be discoverable, governed by rich metadata, and tightly integrated between business and technical layers. He argued that organizations must move away from the illusion of a single version of truth, instead embracing multiple semantic layers tailored to different business contexts while maintaining reconciliation to common data sources. Stephen also highlighted the critical role of metadata, calling it the “engine” that drives trust, transparency, lineage, and interoperability across an enterprise’s data ecosystem.
Debarag complemented these ideas by describing L&T Finance’s journey toward becoming a data-driven organization. He detailed their hybrid architecture built on Google Cloud’s BigQuery and GCS, integrating multiple enterprise systems, AI-driven engines like Cyclops, and their new 360-degree customer analytics platform Nosta Tamus. Governance and data democratization were central to his vision, with strict role-based access to protect sensitive information while enabling over 35,000 employees to use data responsibly. Debarag introduced the concept of “Return on Data Investment” (RoDI) to measure the business value generated by specific data sources—showing, for instance, that alternative data such as payment fingerprints sometimes outperform traditional credit metrics in predicting loan defaults.
In the second half of the webinar, Stephen demonstrated how generative AI and agentic automation can revolutionize data product development by generating semantic mappings, quality rules, and metadata automatically. He stressed that metadata context—not just data—is the key to precision and future-proofing AI systems, enabling interoperability across clouds, databases, and large language models. The discussion concluded with reflections on governance, FAIR principles (Findable, Accessible, Interoperable, Reusable), and the balance between democratization and data protection. Overall, the webinar reinforced that the fusion of AI and data product thinking is essential to building resilient, intelligent enterprises where AI empowers data—and data, in turn, empowers AI.
