Breaking Down Data Silos — and the Culture That Created Them


Original broadcast 10/15/25

For all the excitement around AI, most federal leaders agree on one challenge: their data isn’t ready. Siloed systems, competing priorities, and cultural barriers continue to block progress. But according to Taka Ariga, former Chief Data and AI Officer at the Office of Personnel Management, and Patrick McGarry, Federal Chief Data Officer at ServiceNow, agencies can fix that — if they focus on people as much as technology.

“The number one issue I’ve seen,” says Ariga, “is conflating data management with AI. They’re related, but not the same. You can’t treat your AI program like your IT program and expect it to work.”

Screenshot 2025-10-01 at 8.59.51 PMAriga, who previously served as Chief Data Scientist at the Government Accountability Office, explains that AI requires a new mindset — one that accounts for unstructured, contextual, and evolving data. At OPM, his team introduced “Agile Governance,” bringing legal, privacy, procurement, HR, and technology leaders together in small, recurring sessions to make decisions quickly.

“We limited participation to ten people,” he recalls. “We met every two weeks and made incremental decisions instead of waiting for quarterly reviews. That pace kept innovation ahead of bureaucracy.”

McGarry agrees. He describes data challenges as “socio-technical,” meaning they’re as much about collaboration as code. “Everyone knows we have tons of data,” he says. “The problem is, those who own it don’t always work together. Until you break that cultural fragmentation, your data won’t support your mission.”

He argues that agencies must build governance structures that are living systems — not static PDFs on a SharePoint site. “Governance should evolve as your data does,” McGarry notes. “That’s how you stay aligned with your mission and keep your teams accountable.”

Screenshot 2025-10-01 at 9.00.05 PMBoth leaders emphasize that data control doesn’t mean perfection. Ariga warns against “analysis paralysis” — spending so much time cleaning and structuring data that agencies lose sight of their goals. “Focus on the question you’re trying to answer,” he says. “Perfect data is less important than relevant data.”

Together, Ariga and McGarry outline a vision for what data maturity looks like in 2025: connected systems, collaborative teams, and governance that adapts in real time. “When you tie your data to your mission problems,” McGarry says, “you turn information into intelligence. That’s when AI starts working for you.”

Key Takeaways

  • Cultural collaboration is the foundation of AI-ready data.

  • Agile, iterative governance models accelerate innovation.

  • Agencies should focus on mission-driven outcomes, not data perfection.


This interview was part of Intelligent Government: Smart Strategies to Accelerate AI Innovation presented by ServiceNow. Click here to watch the entire program.