North Highland's Health and Human Services team joined state Medicaid leaders at the 2026 State Healthcare IT Connect Summit in New Orleans, where conversations centered on a shared challenge: how to move AI from experimental pilots to operational programs without compromising public trust. Michael Nourollahi, a Manager in North Highland's Tech and Data practice, led two sessions focusing on operationalizing AI governance and AI use in Medicaid initiatives, both of which drew state leaders from across the country grappling with exactly that same challenge.
In the first session, Nourollahi partnered with a Deputy CTO at a State Agency for Health and Family Services (HFS) to demonstrate what governance-first AI adoption looks like in practice. The message was clear: AI adoption is inevitable, but success depends on building the right structures before scaling.The discussions surfaced a pattern too many agencies encounter: shadow AI usage, fragmented governance, unclear ROI expectations, and an inability to balance innovation with risk. Nourollahi and the Deputy CTO emphasized that breaking the pilot-to-program cycle starts with platform-enabled governance and clearly defined business outcomes.
The State HFS agency brought that principle to life with their customized AI Governance platform, a centralized solution that functions as an intake system, risk classification engine, and audit repository. To find the right fit, North Highland evaluated 14+ vendors against 35+ criteria, ultimately selecting a solution built for public sector needs. All AI use cases must define intended outcomes upfront, with automated risk scoring determining the level of governance review required. As the program matures, high-risk applications will require board approval and human-in-the-loop verification, while lower-risk tools will move faster through structured workflows.
During the roundtable session, state leaders across 4 states revealed consistent practices: separate internal and external use case workflows, cross-functional review boards including legal and security teams and required ROI definitions before deployment. One state's results stood out: AI-powered notice quality assurance eliminated 20 contractors while reviewing 100% of notices, up from just 10%. A policy advisor tool cut response times from 6-7 hours to minutes, benchmarked at 95% accuracy against human performance. The takeaway: measure AI against human benchmarks first, then raise the bar as systems mature.
Leaders framed data maturity and trust as foundational requirements: if you don't understand your data, you're not ready for AI. States are integrating data cataloging and governance with AI strategies, recognizing that AI maturity depends on data maturity. Human-in-the-loop design emerged as a best practice, with states building verification checkpoints into automated workflows. Some states generated synthetic data to enable research without PHI exposure, while others negotiated vendor cost savings after approving AI tools that improved efficiency.
From governance frameworks to real-world results, a clear set of principles emerged from the room separating agencies successfully scaling AI from those still stuck in pilot mode.
Structured governance accelerates adoption when built correctly, moving from policy documents to platform-enabled execution.
States achieving results define success metrics, productivity gains, and ROI expectations before deployment.
Agencies are investing in data cataloging and classification as prerequisites for scaling AI responsibly.
States design AI systems with explicit verification points to build trust while delivering efficiency gains.
States are moving from pilots to operational scaling faster than expected, with vendor partnerships critical to success.
In the public sector, AI adoption lives or dies on trust. Agencies that build it through transparent governance, clear accountability, and measurable outcomes are the ones that will scale successfully.
The message from New Orleans was clear: the agencies moving fastest aren't the ones taking the biggest risks, they’re the ones that built the right foundations first. With the right governance, data strategy, and platforms in place, the path from pilots to enterprise transformation is closer than most agencies think.
Our public health team brings decades of firsthand agency experience to help you get there. Ready to move your health and human services agency from AI pilots to enterprise transformation? Let's talk about what governance-first adoption looks like for you.