PCI Government Services CEO Nick Dunn Redefines AI Governance for the Federal Era
For federal leaders, AI governance is no longer a paperwork exercise. It's the control surface that lets you move fast without breaking trust. Nick Dunn, CEO of PCI Government Services, lays out a model that matches the speed and ambiguity of modern AI while staying inside the guardrails of law, policy and oversight.
With more than 18 years in federal contracting and a $100M-plus portfolio across defense and homeland missions, Dunn argues that AI isn't just another IT upgrade. It's probabilistic, adaptive and capable of producing original outputs-so the operating model must change with it.
AI Is Different: From Guardrails to "Constraints by Design"
Cloud and mobile digitized workflows. Cyber hardened them. AI creates outputs that no one explicitly programmed, which makes traditional, after-the-fact compliance too slow. Dunn's point: institutions built for human timescales must now govern systems that operate on machine timescales.
The pivot underway is simple to say and hard to do: turn policy into code. Bake rules, limits and escalation triggers into data pipelines, model interfaces and deployment workflows so oversight happens in real time. For reference points, see the NIST AI Risk Management Framework and OMB's AI policy memo (M-24-10) for agencies building governance muscle.
- NIST AI Risk Management Framework
- OMB M-24-10: Advancing Governance, Innovation, and Risk Management for Agency Use of AI (PDF)
From Human-on-the-Loop to Human-as-Governor
"Human-on-the-loop" works until the loop gets too fast, like in real-time cyber defense. Dunn proposes a shift: humans set objectives, ethical boundaries and escalation triggers, while agents execute within those constraints. Think architect, not joystick operator.
The hidden risk is institutional amnesia. If a machine made a decision six months ago, can your team still explain why? PCI's answer is an evidence trail-a reconstructable chain from action back to human intent, legal basis and policy. If you can't trace the why, you don't have bounded autonomy. You have opaque risk.
The Readiness Gap Slowing AI at Scale
Agencies face pressure to deploy AI, but many skip the prerequisites. Dunn calls out three gaps that stall programs and erode trust.
- Data readiness: Not just clean data-documented lineage, purpose limits and legal authorization for training and inference. Reporting-grade data often fails those tests.
- Governance foundations: Policies on paper aren't enough. You need roles, accountability, escalation paths and monitoring that actually run in production.
- Leadership enablement: Executives must know what they're authorizing. They need the right questions, the right thresholds and a view of "good" before committing funds.
Principle: trust before speed. Leaders who invest here move faster later-because they aren't fixing fundamentals mid-flight.
Procurement Reality: Consolidation And The "Valley" From Pilot to Production
Best-in-class vehicles and large IDIQs promise efficiency, but they can split the market. Prototypes flow through OTAs; production lives elsewhere. The result is a valley where good ideas stall with no clear bridge to an operational contract.
Another snag: commercial-first only works if solutions integrate. Non-traditionals often sub to primes to clear security and compliance hurdles, and the "innovation premium" gets burned on friction instead of mission impact.
How PCI Federal Is Responding
PCI Federal positions itself as the bridge. As a tribally owned organization built for long horizons, the company serves as the integration layer-bringing proven commercial capabilities into secure, compliant production environments without turning them into isolated islands.
Dunn isn't against consolidation; he's for connectedness. A healthy industrial base needs competition from pilot to platform, with a working bridge in between. The goal for 2026: make that bridge real-policy, contracts, ATO, integration and sustainment tied together.
What Federal Leaders Can Do Now
- Codify governance: Policy-as-code, gated pipelines, automated controls and immutable logs for every deployment.
- Demand traceability: Require lineage, purpose limitation and authorization metadata in RFPs and acceptances.
- Clarify roles: Name accountable owners (CAIO, AO, data stewards) and define escalation triggers before go-live.
- Fund the bridge: Budget for integration, ATO, red-teaming and sustainment-don't stop at the prototype.
- Measure continuance assurance: Track explainability coverage, incident response time, model drift and auditability.
Who Is Nick Dunn?
Nick Dunn is CEO of PCI Government Services and a federal contracting executive with 18+ years supporting U.S. agencies. He oversees a $100M-plus portfolio and previously founded Integrated Federal Solutions, scaling it into a $100M-plus 8(a) prime. His background includes Army contracts legal work and advisory roles in pricing, compliance and risk.
About PCI Government Services
PCI Government Services is part of the PCI Federal enterprise, a tribally owned organization delivering professional services and IT solutions to U.S. agencies. The team focuses on governance, compliance and enterprise IT-helping agencies operationalize innovation inside secure, mission-ready environments.
Events and Resources
Interested in policy, acquisition and operational realities driving AI across government? The Potomac Officers Club's 2026 AI Summit is scheduled for March 18.
If your team needs practical upskilling for AI governance and adoption, explore curated programs by role here: AI courses by job.
Your membership also unlocks:
