White Paper: An AI-Enabled Knowledge Layer for HHS Evidence

HHS holds extraordinary data and evidence assets: clinical records, research findings, claims data, and public health surveillance. But these assets live in separate systems with inconsistent standards, which makes evidence hard to share, validate, and apply. The result is slower science, fragmented decision-making, and missed opportunities to improve health outcomes.

Our white paper makes the case for an AI-enabled knowledge infrastructure as the connective layer HHS needs. The operating model is simple: connect approved HHS knowledge sources, use AI to retrieve and summarize relevant evidence with source traceability, and require human review, governance controls, and audit logs before outputs inform research, policy, operations, clinical guidance, or public communication.

The paper walks through the full evidence pipeline, from generation and synthesis through management, dissemination, and use, and identifies evidence management as the emerging layer HHS does not yet have at department scale. It examines the cross-cutting challenges any agency navigating AI adoption must plan for: interoperability, governance, security, and workforce readiness.

Based on our analysis, we recommend that HHS focus on four priorities:

• Build it right. Establish a named evidence management layer and shared interoperability standards that allow AI to operate safely across HHS data assets.
• Govern it responsibly. Treat privacy and security as ongoing lifecycle processes and establish shared frameworks for validating AI-enabled tools before deployment.
• Grow with it deliberately. Invest in living AI literacy infrastructure that keeps the workforce current as tools and policies evolve.
• Test high-value use cases. Pilot agentic AI systems for living systematic reviews to demonstrate measurable, scalable benefits while preserving human review.

Conselara works at the intersection of health evidence infrastructure, AI governance, interoperability, and workforce enablement for federal health agencies. We built and operate production AI knowledge systems on this exact model, and we welcome the conversation about what a governed AI evidence management pilot could look like for your program.

Download the white paper (PDF)