Automatic Creation of a Semantic Layer for Enterprise Data: The Foundation for Making Your Data AI Ready
A Talk by Sofus Macskassy (Co-founder, Chief Scientist, zaimler)
About this Talk
In today's world, many enterprises suffer from having a large heterogeneous data ecosystem that is siloed and not well understood from the data practitioners who are meant to develop data products that power the enterprise.
This is leading to significant time lost, revenue loss and poor performing models that can hurt their brand. Recent public articles from data leaders suggest that 80% of AI failures are due to data issues and companies are losing billions of dollars due to this challenge.
The key problem that data practitioners are facing is that they do not have a solid semantic layer of their data lake to help them find the right data or understand the data and its intended use.
Additionally, with the advent of LLM, many enterprises now want to use not only their structured data but also their unstructured data. Having a solid semantic layer that glues together these different types of data in a meaningful way is critical for the future of enterprises.
I will in this talk discuss a more holistic approach to creating a semantic layer that directly addresses these pain points. I will describe innovative approaches to automate a significant portion of ontology creation and knowledge graph building from an enterprise's own data lake.
I will ground the value of this in specific use cases to show how this semantic layer will address many of the key challenges in creating trustworthy data that is AI ready and how this will power data products such as RAGs.
About The Speakers
We’re a cloud tech company that provides organisations around the world with computing infrastructure and software to help them innovate, unlock efficiencies and become more effective. We also created the world’s first – and only – autonomous database to help organise and secure our customers’ data.
