What happens when health systems stop waiting on vendor roadmaps and start building AI workflows themselves? At Noteworthy 2026, Notable’s Gabriella Rozenblat talked with three leaders who have launched Builder programs and built AI workflows themselves in Notable’s Flow Builder:
- Brian Schuetz, Executive Director – MIT Health
- Jeffrey Pan, Senior Director of Digital Health, Population Health, and AI – UC San Diego Health
- Dr. Todd Beardman, CMIO – NKC Health
The panelists described a Builder as someone who understands both the technology and the operational problem it solves. They shared how they staff Builder teams, including Jeff’s advice to “always [hire] two.” They also explained how to choose what to build and how to measure success. Brian described a new kind of return he calls “non-Newtonian ROI”: work that simply wasn’t possible before AI agents, such as a three-year retrospective review of incidental imaging findings to check whether each one was followed up.
Lessons you’ll learn in this session:
- Hire Builders who understand both the problem and the technology. Some health systems use senior staff because the problems they’re solving are enterprise-wide.
- Hire in pairs. Pair one technical Builder with one who knows workflows, so they can design and troubleshoot together.
- Give new Builders a real problem they can finish in a week. Early wins build confidence faster than training alone.
- Use the Builder Community. Builder cohorts, shared templates, the Builder Slack channel, and building together on calls help new teams get unstuck.
- Measure all three kinds of ROI. Count hard ROI, soft ROI, and the value of work that wasn’t possible before.
- Define the full life cycle before you build. Decide who builds, validates, maintains, and owns each workflow. As Brian Schuetz said, “building is the easy part.”
Watch the full panel to learn how health systems are building in-house capacity to scale AI automation across the enterprise.





