Establishing a Scalable, Trustworthy AI Delivery Model for Veteran Health Care

The Challenge

VA recognized the potential of AI to transform care delivery and reduce administrative burden, but struggled to move AI initiatives from research into real-world clinical and operational pilots with measurable value. While promising ideas emerged across VA's Veterans Integrated Service Networks (VISNs) and programs, the NAII lacked the structured delivery capacity to translate these concepts into trusted, validated solutions.

The challenges were systemic. Clinicians spent excessive time on manual, document-heavy workflows — reviewing and routing community care documents, creating discharge instructions, and managing administrative tasks that kept them from patient care. Hospital staff saw Veterans struggle with discharge instructions, leading to avoidable follow-up issues. Yet without structured discovery, value measurement, and pilot design processes, NAII couldn't confidently identify which AI solutions would deliver real impact or gain user adoption.

The complexity was multilayered: health care AI in a highly regulated federal environment with strict requirements under the Health Insurance Portability and Accountability Act (HIPAA), protected health information (PHI), National Data Systems (NDS), VA Handbook 6500, and VA’s trustworthy AI framework. Hybrid funding constraints meant limited ability to write code directly. Distributed infrastructure, such as centralized clinical data versus site-local imaging data, complicated technical implementation. And every solution required navigating VA badging, background investigations, and system access while operating under a fixed sprint cadence and potential government shutdown risks.

For VA leadership, the stakes were clear: without validated evidence of value, safety, and adoption, AI solutions couldn't scale. For clinicians already facing burnout, poorly designed AI tools risked adding complexity rather than relief.

Our Approach

Aquia Nava II LCC, our joint venture with Nava Public Benefit Corporation, embedded a delivery-first, human-centered AI enablement model directly into NAII, providing the multidisciplinary structure and rigor needed to mature AI use cases from concept through validated pilots. We coordinated a multi-vendor, cross-functional team that aligned tightly with NAII's lifecycle: Discovery → Solutioning → Piloting → Evaluation.

We mobilized within 10 days of award, launching discovery planning and use-case alignment within the first month. Even before gaining personally identifiable information (PII) or PHI access, the team launched discovery planning and use-case alignment work, maintaining momentum through two-week agile sprints with VA-approved goals and deliverables.

The approach integrated human-centered design with value-driven product management. We conducted ethnographic research and clinician interviews to map actual workflows, ensuring AI solutions aligned with real operational needs rather than just technical feasibility. For every use case, baseline metrics, value hypotheses, and acceptance criteria were developed to measure ROI before committing to full pilots.

Our team brought deep federal health and VA domain experience across program management, product strategy, UX research, AI/ML, solution architecture, and security compliance — disciplines that NAII had limited internal capacity to coordinate. This multidisciplinary integration will enable us to assess build-versus-buy approaches, configure low- and no-code AI solutions for evaluation, and navigate VA's governance requirements as a unified delivery motion.

Technical work focused on AI-assisted capabilities that addressed documented pain points: document ingestion and summarization to reduce clinician review time, discharge documentation optimization to improve Veteran understanding, and early technical discovery for imaging-based AI under VA's distributed infrastructure constraints.

The Impact

The engagement established what NAII needed most: a repeatable, scalable model for AI delivery and intake. High-priority use cases moved from abstract ideas into structured discovery with validated workflows, reducing the risk of deploying AI solutions that clinicians wouldn't adopt or that would fail to deliver measurable value.

The work supported multiple concurrent AI product lanes, with dozens of stakeholder and clinician interviews conducted across VISNs. Standardized sprint cadence and reporting gave leadership clearer visibility into value, risk, and readiness for scaling — enabling confident investment decisions backed by evidence rather than assumptions.

The rigor and structure proved that VA could responsibly advance AI adoption at enterprise scale while maintaining the trustworthiness Veterans and clinicians require.

Most significantly, the engagement shifted NAII's trajectory from fragmented experimentation to strategic, human-centered delivery. VA can now validate AI solutions against real clinician workflows before full deployment, align AI tools with documented operational needs, and make evidence-based scaling decisions — all within the governance framework that protects Veterans and ensures responsible AI use.

The foundation is set for AI that enhances human capability rather than replacing it: reducing clinician burnout, improving care coordination and discharge outcomes, and increasing Veteran trust in digital health tools. With a proven delivery model in place, NAII can confidently pursue AI initiatives that serve Veterans faster and more reliably.

Aquia Nava II LLC partnered with the Department of Veterans Affairs National Artificial Intelligence Institute under the VA Secure, Performant, Reliable, and User-Centered Experiences (SPRUCE) contract vehicle to build trustworthy AI delivery capacity within the Veterans Health Administration.