Build Your AI Strategy Around Data, Not Infrastructure

Most companies building AI strategies are starting in the wrong place. They focus on infrastructure — cloud, compute, deployment architecture — and treat data as something to be dealt with later. According to Juan Orlandini, CTO of North America for Insight Enterprises, that's backwards.

The Real Problem: Siloed Data

Orlandini opens with a striking example: a CIO who employed 800 data scientists spread across business units that didn't collaborate. Each had valuable data, but siloed communication prevented them from solving each other's problems. When Insight helped establish an AI Center of Excellence, teams began sharing knowledge — and solved challenges that had stumped them for years.

Key Takeaways

• Data is the new oil — but only if you can access it. LLMs are powerful, but you can't extract value by pointing them at incompatible data lakes with conflicting security controls.
• A data-centric model enables AI to access multiple repositories without normalizing every dataset first. This unlocks value across the data estate.
• Don't try to transform everything at once. Successful AI transformations start small, prove value, and accelerate.
• Infrastructure follows data, not the other way around. Where your data lives matters less than making it accessible with proper security controls intact.

🔗 Read the full article on Fast Company