Overview on Data Foundations for Generative AI
Gartner study found 45% of execs piloting GenAI, but only 10% have 1st app in prod; <10% PoC to production. AWS CDO survey cited data quality as #1 barrier. 82.2% of Fortune 1000 increasing data & analytics for AI. Session aims to help partners guide customers on data foundations to improve AI investment success.
The show focused on the critical but often overlooked aspect of data foundations in generative AI adoption. According to Gartner statistics referenced in the discussion, fewer than 10% of generative AI projects successfully transition from proof of concept to production, largely due to inadequate data foundations. The panel emphasized that while many organizations focus primarily on the AI models and outcomes, the underlying data infrastructure, governance, and quality controls are essential for successful implementation.
The experts outlined several key components of data foundations, including infrastructure for centralizing data ecosystems, modernization stages for preparing data, and purpose-built databases with vector embedding capabilities. They stressed that organizations need to consider factors like data cleansing, standardization, and governance before implementing generative AI solutions. The discussion highlighted AWS services like S3, Lake Formation, Glue, and Data Zone that can help build these foundations.
The session included a detailed demonstration of a clinical trials bot showcasing how proper data foundations enable efficient and accurate AI applications. The demo showed how unstructured clinical trial data could be transformed into a queryable system using vector embeddings, delivering relevant results in milliseconds from a database of half a million records, while maintaining transparency about data sources to build user trust.
Key Highlights:
- Data foundations include infrastructure, governance, processes, and quality controls that make data usable for AI
- Organizations should build production-ready systems from day zero rather than focusing solely on POCs
- Lambda functions play a crucial role in building practical generative AI solutions
- Vector databases and embedding models are essential components for efficient data retrieval
- Proper data foundations can lead to 51% higher return on AI investments
- Success requires both technical implementation and clear business value alignment
- Responsible AI principles should be incorporated from the beginning, including transparency about data sources
Hosts of the show 🎤
Ibrahim Emara - Database Specialist SA @AWS
Guests 🎤
Yadukishore T - Sr Solutions Architect : Data,Analytics and Generative AI @ AWS
Martin Pomykala - Global Business Development + Partner Strategy @ AWS
Eric Miller - Vice President of Data & AI @ ClearScale
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article