This interview was filmed on location at The Helix, Booz Allen’s Center for Innovation in Washington, D.C., as part of the event DE25: Driving Outcomes through Data. The program features top technology leaders from the public and private sectors sharing insights on cloud transformation, agentic AI, fraud prevention, and data governance. Through a series of dynamic conversations, the program captures how agencies are aligning digital infrastructure with mission needs to deliver real results for the American people. Watch the full show.
Data has become the federal government’s most valuable asset—and one of its most complex. In the “Digital Transformation: Improving Government Efficiency through Data” segment of Driving Outcomes through Data, three public and private sector leaders share how agencies are navigating the operational, cultural, and technical challenges of data-driven transformation.
ss.
“You can have all the data in the world, but without the right structure, governance, and access, you’re just building a swamp,” Carroll warns.
The Foundation: Classify, Govern, Monitor
Matt Carroll opens the conversation by laying out three critical components that agencies must put in place to move from raw data accumulation to meaningful data utilization: classification, approval, and monitoring.
He points out that while agencies in the intelligence community have long operated with strict data classification systems, many civilian agencies are still working to develop shared understanding of sensitivity levels, access rights, and risk tolerance.
“It’s not very sexy, but classification is where it starts,” Carroll says. “If you don’t know what your data is, you can’t control it or use it effectively.”
He also underscores the need for clear approval workflows, especially in an environment where technical barriers to access are falling. As users gain the ability to query data without writing code, agencies must formalize who can grant access, under what conditions, and with what safeguards.
Finally, agencies must have the ability to dynamically monitor data usage—not just to ensure compliance, but to remove access when it’s no longer needed and prevent misuse in real time.
The Architecture: From Pipelines to Semantics
Dan Tucker picks up the thread by walking through the modern data stack—from pipelines and metadata to observability and access control. While many organizations have successfully built the technical components to move and manage data, Tucker points out that a critical layer is often still missing: semantics.
“You can have your data pipelines and your dashboards,” he says, “but without a semantic layer, your users don’t really understand what the data means.”
The semantic layer acts as a translator between raw data and mission needs, providing common definitions, relationships, and business logic that can be reused across systems and teams. This layer becomes even more important as agencies adopt AI and machine learning, where understanding the context of data is essential to generating reliable insights.
Tucker also highlights the importance of decentralized governance combined with centralized visibility—giving mission teams the flexibility to control their own data domains while still supporting enterprise-wide interoperability.
The Playbook: Strategy, Architecture, Governance, Services
Austin Gerig brings the agency perspective, describing how the SEC has operationalized the responsibilities outlined in the Foundations for Evidence-Based Policymaking Act, commonly known as the Evidence Act. Gerig and his team built the SEC’s data office using a four-part functional model: data strategy and use, data architecture, data governance, and data services.
Each function is staffed with dedicated leads and supported by cross-agency governance committees to ensure buy-in and adoption. Gerig emphasizes that while every agency has unique needs, this structure maps well to the statutory requirements and 
mission demands across government.
“We wanted to build an organization that wasn’t just compliant—but also effective,” Gerig explains.
As an early proving ground, the SEC focused its data strategy on a high-impact mission area: insider trading detection. This initiative brought together analysts, technologists, and data experts to work as a unified team, applying the agency’s data strategy to a real-world problem.
The result? The largest front-running case ever detected by the SEC, powered by analytics, infrastructure, and governance working in sync.
Cross-Sector Parallels and Lessons Learned
One of the most striking takeaways from the panel is the consistency between public and private sector data practices. Carroll notes that whether a team is running clinical trials or enforcing market rules, the requirements for data control, access, and accountability are strikingly similar.
Tucker adds that the biggest challenge isn’t technical—it’s scale. In both domains, organizations must figure out how to equip domain experts with the tools and context they need, without creating bottlenecks or overburdening a central team.
That’s where AI and automation come in. As data sets grow and the demand for access expands, agencies must modernize both workforce roles and toolsets to enable secure, scalable decision-making.
“Access to data is not just a technical challenge,” Tucker says. “It’s a team sport.”
The Road Ahead: Semantic Infrastructure and Agentic Systems
Looking forward, the panelists agree that agencies must focus on semantic infrastructure—not just to support current analytics, but to enable agentic AI systems that can reason, act, and adapt based on mission context.
Gerig points out that a solid semantic foundation is already paying dividends at the SEC, where the team’s architecture can now support any structured dataset across the agency.
“You have to start with where you want to go,” he says. “If you build the right strategy and structure, the rest can follow.”
As federal agencies look to modernize and future-proof their missions, this segment underscores a simple but powerful truth: data only delivers outcomes when it’s governed, structured, and used with purpose.