Hi Jan,

Every copy of your data is a governance obligation. A pipeline dumps customer records into a staging table. The staging table sits ungoverned for months. An AI agent with broad access finds it, and queries it on behalf of 50 users who were never supposed to see unmasked PII.

This is the silo problem, and it compounds. The book identifies five ways silos actively destroy AI analytics: ETL lag, conflicting logic across copies, schema drift, governance fragmentation, and exploding security surface area.

The architectural answer is query federation, querying data in place, across cloud warehouses, data lakes, SaaS apps, and on-premise systems, without moving it. Zero copies. One governance layer. Sub-second responses.

P.S. - Join Alex Merced for a live 60-minute session: Your AI Agent Can’t Answer That Question, Here’s Why and How to Fix It on Thursday, April 2 to learn more about the book.

This email was sent to nie@niepodam.pl. If you’d rather not receive this kind of email, you can unsubscribe.

610 Brazos St., Suite 300 Austin, Texas 78701, US