Problem Statement
Federal agencies today are under mounting pressure to modernize legacy systems, integrate emerging technologies, particularly artificial intelligence (AI), in order to transform operations, enhance decision-making, and deliver more effective services. At the same time, they must maintain rigorous compliance, security, and transparency requirements. In addition, AI modernization requires significant cultural change. Employees must adapt to new workflows and mindsets. Traditional Centers of Excellence (COEs) often struggle to support this, while a Federated COE embeds SMEs into delivery teams, enabling hands-on engagement and trust-building within each program.
Traditional centralized Centers of Excellence (COEs), while designed to promote consistency and standardization, often fall short in enabling agile innovation and delivering scalable, mission-aligned impact across diverse federal programs. These centralized COEs tend to struggle with slow responsiveness to field-level needs, limited visibility into actual project conditions, and bottlenecks in governance workflows. Multiple state IT consolidation efforts have revealed similar patterns. The National Association of State CIOs (NASCIO) reported that although centralization sometimes achieves infrastructure cost savings, it often leads to bureaucratic bottlenecks, hidden costs, and reduced agency flexibility that hinder technology adoption and responsiveness (NASCIO, 2022).
As a result, the adoption of new tools and practices is frequently hindered, leaving agencies unable to capitalize on innovation opportunities that could strengthen mission delivery fully. AI implementation presents unique challenges for federal agencies, including rapidly evolving technologies, talent shortages, fragmented pilot efforts, and difficulty transitioning promising proofs-of-concept into full-scale enterprise adoption. Traditional centralized COEs often lack the program-level context needed to successfully scale AI solutions across diverse mission areas.
In large federal organizations, where programs vary significantly in mission objectives, operational maturity, and resource constraints, centralized COE models often generate guidance that lacks applicability to field conditions. This disconnect leads to underutilized resources and missed opportunities for innovation, leaving agencies with persistent modernization challenges. While centralized COEs benefit from consolidated expertise, clear standards, and economies of scale, they can also introduce rigidity and slow adoption. A federated model, by contrast, retains centralized governance and knowledge sharing while enabling greater responsiveness and mission alignment at the program level.
Proposed Solution
To overcome the limitations of centralized COEs, especially in the fast-moving domain of AI, federal agencies should adopt a Federated COE model. This model preserves the governance, standardization, and oversight benefits of a centralized mission while decentralizing execution.
It does so by embedding subject matter experts (SMEs) within active delivery teams across programs. Federal efforts, such as the GSA’s COE initiative, have demonstrated the power of embedding digital experts directly into agency teams. GSA describes its COEs as “able to change or grow to fit the needs of different sectors of government,” highlighting the importance of tailoring innovation support to each program’s specific context (GSA COE Program Overview, 2020)
Embedded SMEs maintain a formal reporting line to the central COE while working day-to-day within delivery programs through a dotted-line relationship. This structure ensures that while SMEs operate within and contribute to program teams, they maintain alignment with enterprise standards, evolving policy, and cross-program knowledge transfer.
These forward-deployed SMEs devote a portion of their time to COE functions, which include surfacing real-world innovations, validating tools and processes in live delivery contexts, sharing lessons learned horizontally across teams, and promoting lightweight standards grounded in proven field success.
By aligning COE expertise directly with program delivery, agencies can create an adaptive framework that accelerates technology adoption while ensuring innovations remain aligned with mission needs and operational realities. The Federated COE model offers an effective mechanism for agencies to modernize their systems while maintaining a balance between central oversight and program-level autonomy.
This federated approach mirrors recent evolutions in the Department of Defense. The GAO found that DoD’s Joint Artificial Intelligence Center (JAIC), now part of the Chief Digital and Artificial Intelligence Office (CDAO), structured initially as a highly centralized AI COE, evolved to focus on providing technical services, acquisition support, expertise, and best practices to embedded AI efforts across the Military Services and Combatant Commands—enabling more agile, mission-aligned AI adoption (GAO-22-104670, 2022)
Description of the Federated COE Model
The Federated COE model enables mission-aligned innovation by positioning SMEs where work occurs, ensuring innovations directly address operational requirements. Through their embedded roles, these SMEs continuously identify promising use cases, validate emerging tools, and disseminate best practices across programs through communities of practice. This accelerates the pace of technology adoption, minimizes the gap between pilot efforts and enterprise-wide implementation, and fosters broad peer acceptance.
Furthermore, by leveraging scarce expertise more efficiently, the Federated COE allows specialized resources in areas such as AI, DevSecOps, cloud computing, and Zero Trust security to exert influence across multiple programs without being tied to a single project or confined within a central team. As innovations emerge organically from within delivery teams, peer-to-peer buy-in increases, driving higher adoption rates than top-down mandates often achieve. Simultaneously, guardrails and governance remain in place to ensure that standards and security protocols are consistently met while allowing flexibility in execution.
AI/ML Enablement in Government
The Federated COE model is particularly well-suited to emerging technology domains, such as AI and machine learning. Instead of creating a detached, centralized AI COE, agencies embed AI SMEs across programs where they can work directly with delivery teams to identify valuable use cases, such as fraud detection or document summarization. These experts pilot AI tools within real workflows, share results and lessons learned across programs, and coordinate lightly with a central oversight office that ensures compliance and security requirements are maintained. This federated approach enables faster deployment, higher success rates, and more efficient use of AI funding by grounding innovation in practical, mission-driven needs.
Implementation Considerations
The successful implementation of a Federated COE model requires a clear definition of SME roles and expectations, strong feedback loops between program teams and central leadership, and structured communities of practice to support horizontal knowledge sharing. Ultimately, leadership support is crucial in striking a balance between decentralized autonomy and organizational accountability, ensuring that federated activities remain aligned with agency goals and regulatory requirements.
Successful implementation requires careful management of risks. Federated SMEs must avoid becoming isolated mini-silos, and periodic rotations or cross-program working groups can ensure continuous knowledge sharing. Leadership should also formalize SME roles (e.g., allocating 5–10% of their time to COE functions), supported by regular reporting and centralized dashboards that monitor adoption, compliance, and performance outcomes.
Rotating SMEs between program assignments and central enterprise roles helps reinforce a virtuous cycle: field experience enhances enterprise-level guidance, while exposure to central strategy enables SMEs to drive more mission-aligned innovation in future deployments.
Recommended Path Forward
The Federated COE model offers federal agencies a pragmatic and scalable approach to modernization. By combining centralized vision and decentralized execution, agencies can remain agile, mission-focused, and better positioned to meet the dynamic demands of modern federal operations. This approach enables agencies to accelerate AI adoption, promote continuous learning across programs, and foster innovation without sacrificing oversight or governance.
Learn More
If you found this white paper helpful and would like more information, please contact KGSinfo@koniag-gs.com. In your email, be sure to reference the white paper and include any specific follow-up requests or details you’d like us to provide.