Future-ready federal AI: Meeting new OMB guidelines with Domino

Domino2025-04-29 | 7 min read

Return to blog home

OMB M-25-21: Advancing governance, innovation, and responsible AI use

OMB M-25-21 sets strict requirements to ensure responsible AI governance, use case transparency, and generative AI (GenAI) safety. Here's how Domino directly supports these mandates:

Centralized AI governance and oversight

  • OMB context – Designate CAIOs and establish AI Governance Board: Domino gives chief AI officers (CAIOs) and governance boards enterprise-wide oversight into all AI assets — models, data, risk levels, and compliance status — via a unified platform.
  • Navy and DIU in action: The Navy’s use of Domino provides real-time visibility into all models used for underwater threat detection. DIU program leads can monitor development and deployment activity with full traceability, fulfilling the spirit of CAIO-driven oversight structures.

Transparent AI use case inventories

  • OMB context – Maintain and publish inventories of AI use cases: Domino logs all models, datasets, stakeholders, and purposes — supporting public inventory requirements and internal accountability.
  • Navy and DIU in action: The AI system for sonar signal classification was tracked from development to deployment. The Navy documented the full model lineage and operational use case, making it easy to compile inventory records for reporting and auditing.

Generative AI policy readiness

  • OMB context – Govern the use of GenAI with secure experimentation and access controls: Domino enables safe experimentation with GenAI through policy-enforced access, version-controlled environments, and full auditability.
  • Navy and DIU in action: While the Navy’s use case focused on classical ML, Domino’s deployment included the same governance mechanisms — segregated environments, traceability, and access logging — that are now required for GenAI oversight.

Minimum risk practices for high-impact AI

  • OMB context – Apply risk mitigation like testing, auditing, and bias detection: Domino includes built-in bias detection, reproducibility, audit logging, and validation workflows to meet minimum practice standards.
  • Navy and DIU in action: For mission-critical detection models, Domino’s reproducibility and audit capabilities allowed the Navy to validate model decisions and retrace logic under operational conditions — key to managing AI risk.

Innovation acceleration

  • OMB context – Remove barriers to innovation: Domino accelerates AI workflows by enabling reuse of models, code, and data pipelines — reducing time to deploy while maintaining full governance.
  • Navy and DIU in action: The Navy reused standardized pipelines and containerized environments to speed experimentation. Domino eliminated friction often caused by security reviews or infrastructure bottlenecks, enabling faster iteration.

From static contracts to smart AI acquisition

OMB M-25-22 introduces a modern procurement model that requires performance-based evaluations, interoperability, and clarity on IP and data rights.

Interoperability and avoiding vendor lock-in

  • OMB contextPromote open standards and avoid proprietary lock-in: Domino is cloud-agnostic and integrates with Jupyter, PyTorch, MLflow, and other open and commercial tools — ensuring flexibility and portability.
  • Navy and DIU in action: Domino’s open architecture allowed the Navy to integrate their preferred tooling and infrastructure — both open source and commercial. No vendor lock-in meant freedom to adapt and scale across mission environments, choosing best-in-class tools across the model lifecycle orchestrated by Domino’s platform.

Performance-based AI tracking

  • OMB context – Track performance metrics like accuracy, drift, and fairness: Domino continuously monitors model behavior to ensure performance-based accountability in procurement and operations.
  • Navy and DIU in action: Domino helped the Navy monitor threat detection model performance over time, evaluating accuracy in varied operational conditions. These metrics laid the foundation for performance-based decisions and improvements.

IP and data ownership controls

  • OMB context – Retain control of AI models, data, and workflows: Domino ensures full agency ownership of all artifacts — securely stored, versioned, and traceable.
  • Navy and DIU in action: The Navy retained ownership of all model artifacts, datasets, and outputs. With Domino, they could access historical versions, retrain on new data, and maintain control without third-party barriers.

Cross-functional collaboration

  • OMB context – Involve mission, procurement, and technical teams throughout: Domino provides a secure, shared workspace where interdisciplinary teams can build, test, and govern AI together.
  • Navy and DIU in action: Program managers, mission leads, and data scientists across multiple contractors collaborated seamlessly on Domino — sharing updates, running experiments, and maintaining a common system of record for ongoing development and delivery.

Security and compliance built for government

Security is foundational to Domino. Our cloud-native, Kubernetes-based platform is built to minimize risk while enabling agility for modern AI development. With Domino Cloud, agencies get a private, dedicated SaaS environment that continuously monitors code, data, and model activity. Every asset — datasets, models, environments, results — is versioned and traceable for full audit readiness.

We’re trusted where it matters most:

  • Deployed in DoD IL5 environments and into Iron Bank
  • Certified for ISO 27001:2013, SOC2, GDPR, HIPAA
  • Supports 21 CFR Part 11, EudraLex Annex 11, WHO Annex 5

With Domino, agencies can meet today’s toughest compliance standards — while moving at mission speed.

Get ahead of the next mandate

New AI policies will keep coming. With Domino, you won’t have to scramble to meet them. Whether you’re a chief AI officer building governance frameworks, a procurement team navigating M-25-22, or a mission lead deploying GenAI safely — Domino gives you the confidence to move forward, securely and at scale.

Let’s build the next generation of government AI — compliant, transparent, and ready for anything. Check out Domino for the public sector for more insights and solution recommendations.


Domino Data Lab empowers the largest AI-driven enterprises to build and operate AI at scale. Domino’s Enterprise AI Platform provides an integrated experience encompassing model development, MLOps, collaboration, and governance. With Domino, global enterprises can develop better medicines, grow more productive crops, develop more competitive products, and more. Founded in 2013, Domino is backed by Sequoia Capital, Coatue Management, NVIDIA, Snowflake, and other leading investors.