When the Massachusetts Innovation Network and the Middlesex 3 Coalition convened leaders from industry, academia, healthcare, and the public sector for a summit on operationalizing AI, the framing was intentional and refreshing. This wasn’t a conversation about what AI could do. It was a working session on what it actually takes to make AI work in organizations, in systems, and in practice.

Uwe Hohgrawe, Professor and Associate Dean of Global Learner Access, Strategic Partnerships, and AI at Northeastern University’s College of Professional Studies, participated in the panel on Healthcare Delivery and Patient Impact, joining colleagues from academic medicine, pharmaceutical industry, life sciences policy, and university-based health innovation. That cross-sector mix was itself instructive. When a table spans the lab, the clinic, the policy office, and the lecture hall, abstractions give way to operational reality quite quickly.

One of the keynote speakers Matt Ferguson, a Master Principal Cloud Architect at Oracle, noted that there are three simple lenses through which to think about operationalizing AI: people, process, and technology. Simple to name, genuinely hard to align.

The progress happening across health innovation is real. AI-enabled tools either already are or can accelerate diagnostics, improving patient outcomes, and creating efficiencies that were unimaginable a decade ago. But the conversation kept returning to the same friction points, and they weren’t primarily technical.

The Obstacles That Aren’t Technical

Legacy systems remain a stubborn obstacle. Even the most sophisticated AI model hits a wall when the infrastructure it needs to integrate with was built for a different era. Updating those systems requires capital, coordination, commitment by leadership, and institutional will that don’t appear on a product roadmap.

Workforce preparedness is the more urgent challenge. Health organizations are asking professionals to adapt to technologies that are evolving faster than training pipelines can keep pace with. However, right now, many people and systems are hesitant and treat AI (e.g., AI-generated documentation) with suspicion, and that caution is understandable. That means that organizations must carefully investing not just in technical literacy but in the broader professional skills such as critical judgment, ethical reasoning, and workflow redesign. The kind of skills that allow people a thoughtful approach to work with AI rather than simply around it.

Trust Is Built Through Building

The point that anchored much of the discussion: technical excellence is not sufficient. A model can perform brilliantly in isolation and still fail entirely in deployment if the people who will use it weren’t part of building it. AI systems in health settings must be developed alongside the life sciences workforce, not above or around it. Trust is not a feature added at launch. It is built through the process of building itself.

Across all three thematic panels of workforce readiness, healthcare delivery, and public service, the summit reinforced a single through-line: operationalizing AI is an organizational challenge as much as a technological one. Measuring ROI means asking what an organization values. Evolving services means deciding who is at the table when those services are redesigned. Managing friction means being honest about where change actually lands and on whom.

The future is not arriving on a schedule anyone controls. But how organizations prepare people, redesign processes, and integrate technology with integrity can be shaped with intention.

About the Panelist

Uwe Hohgrawe is Professor and Associate Dean, Global Learner Access, Strategic Partnerships, and AI. He presented at the Massachusetts Innovation Network and Middlesex 3 Coalition AI Summit as part of the Healthcare Delivery and Patient Impact panel.