GenAI is currently evolving faster than most organizations can adapt. At the same time, most data platforms were not built for GenAI—they were built to power dashboards and batch analytics.
Operationalizing GenAI at scale requires the ability to manage large volumes of unstructured data, low-latency inference, real-time context delivery, retrieval- augmented generation (RAG), and prompt orchestration.
DBTA recently held a webinar, Modern Data Platforms for Operationalizing Generative AI at Scale, featuring David Millman, lead partner sales engineer, System Integrators and GSIs, Fivetran and Randy Rouse, field CTO, Quest Software who discussed key requirements for modern data platforms along with best practices for integrating vector databases, feature stores, retrieval systems, and more.
Millman asked, “What if you could take a repetitive 6–8-week Data Engineering task and accomplish it in less than a week?” AI and automation can help achieve that and that was the problem posed to Fivetran while working with Saks Fifth Avenue.
However, there were challenges in operationalizing AI at scale. This included:
- Scaling data operations were costly and slow - Data operations required 4-5X more staff; Saks struggled to adapt quickly, limiting personalization and the ability to deliver luxury ecommerce experiences at speed.
- Cumbersome pipelines slowed AI adoption - Custom ETL pipelines lacked monitoring and were slow and resource intensive, requiring constant engineering maintenance and delaying data-driven decisions.
- Siloed customer data blocked AI-driven insights - Customer data across supply chain, ecommerce, CRM, loyalty, and marketing systems existed in silos, forcing manual reporting and delaying campaign launches.
Through their partnership, Fivetran powers Saks’ modern data foundation, Millman said. With a modern, cloud-based data stack, Saks freed teams to focus on in-depth analysis and explore new technologies like AI, ML, and GenAI.
LLMs and real-time NLP tools to provide service agents with data-driven recommendations and actionable insights. Standardized reporting across the enterprise and brand partners enabled faster decision-making and greater data utilization.
And Saks improved customer relationships by tailoring campaigns, product recommendations, and personalized experiences for each customer.
“This is the beauty of the data ecosystem we’ve built with Fivetran as a core piece: We can think about data in a fundamentally different way, less consumed with the cost of getting it in and instead focusing on the value it can bring to our customers and brand partners,” said Mike Hite, CTO, Saks.
AI is disrupting the traditional data management landscape and as a result, the market is “converging,” explained Rouse. With a data management platform, data teams can benefit from the reduced operational complexity, drive data engineering productivity, save on license costs, and improve ROI.
The Quest erwin Data Management Platform offers 5 integrated components working together to deliver raw data into a strategic, reusable business-aligned resource for cross-organization collaboration. It provides real-time measurements across nine categories, providing quantifiable trust metrics for data assets. And the intelligent automation of data modeling tasks and stewardship tasks, reduces manual overhead while improving data accuracy, Rouse said.
erwin helps “our customers build modern, trusted foundations for AI with reduced risk, greater automation, and increased agility,” Rouse concluded.
For the full webinar, featuring a more in-depth discussion, Q&A, demo, and more, you can view an archived version of the webinar here.