Everyone wants better AI models. A few days back at Data Citizens on the Road by Collibra, I sat down with Reece Griffiths, Field CTO at Collibra on The Ravit Show, to discuss one of the biggest challenges facing enterprise AI today: unstructured data.
For years, data governance focused primarily on structured data.
But AI is changing the game.
Today, enterprise knowledge lives across PDFs, presentations, images, documents, emails, and shared drives. If that content isn't properly governed, AI systems can quickly run into problems:
* Generating answers from outdated or draft documents
* Exposing sensitive information due to missing confidentiality labels
* Missing relevant content because of poor metadata and classification
One concept from our discussion really stood out:
Knowledge decay.
Even the most advanced AI models will struggle if the underlying knowledge base is stale, incomplete, or poorly maintained.
We also discussed why enterprises are moving toward unified semantic models that connect structured and unstructured data, allowing AI systems to understand business context consistently across the organization.
The takeaway?
The future of enterprise AI won't be determined solely by model performance.
It will be determined by the quality, freshness, and governance of the data behind it.
#Data #DataCitizens #Collibra #AI #GenerativeAI #DataGovernance #AIGovernance #EnterpriseAI #Metadata #DataManagement #TheRavitShow
