Luma Health co-founder Aditya Bansod argues that a new coordination layer called headless AI could finally solve the manual bottlenecks that still plague hospital operations despite years of technology investment
Healthcare organisations have spent enormous sums on AI tools over the past several years. Documentation assistants, clinical decision support systems, patient messaging platforms, scheduling automation. The investment has been real and the results measurable. But there is a problem that none of those individual applications solve, and it is the one that still dominates the daily experience of healthcare workers: moving information manually between systems that do not communicate with each other.
That is the problem Aditya Bansod, co-founder and president of Luma Health, believes the industry needs to tackle next.
What Headless AI Actually Means
Bansod describes headless AI as a coordination layer that works across applications rather than inside any single one. Instead of requiring an employee to navigate between an EHR, a payer portal, a fax queue, and a scheduling platform to complete a prior authorisation, a headless AI system would reach across all of those systems simultaneously, gather the relevant information, reason over it, and push the result back into the appropriate workflow.
The term headless refers to the absence of a user interface. The AI does not live inside any one application. It coordinates across all of them.
Building on Existing Investment
Critically, Bansod frames this not as a replacement for the robotic process automation, workflow engines, and AI assistants that health systems have already deployed but as a way of connecting them. The goal is to stitch together investments that are already working individually but cannot yet address problems that span multiple systems.
The Governance Question
As AI starts taking on tasks by itself throughout enterprise systems, Bansod says governance has to come first, not last. Each AI agent needs its own identity, specific permissions, and role-based access—plus, there has to be a record of everything they do. Application and platform audit trails both matter, so compliance teams can see exactly what happened and why.
He puts it simply: healthcare should focus on treating patients, not endless administrative tasks. Let AI handle the coordination, and give doctors and staff the freedom to focus on real care.
