Healthcare’s Agent Bytecode: X12 as the Harness for AI Agents — Vasant Kearney, Onlay
Healthcare agents can use X12 transactions as a typed contract across calls, portals, and claims. The standard constrains actions and validation, but payer data can still be stale or wrong.
The talk maps insurance work onto public X12 transactions: **270** for eligibility requests, **275** for supporting information, **999** for syntax acknowledgment, and **835** for payment details. The same contract can ground phone, browser, desktop, and EHR actions.
Treat a stable domain protocol as agent bytecode: validate generated actions against allowed fields, permissions, and lifecycle states. Preserve multimodal evidence when reducing it to text would discard context needed later.
The talk maps insurance work onto public X12 transactions: **270** for eligibility requests, **275** for supporting information, **999** for syntax acknowledgment, and **835** for payment details. The same contract can ground phone, browser, desktop, and EHR actions. Treat a stable domain protocol as agent bytecode: validate generated actions against allowed fields, permissions, and lifecycle states. Preserve multimodal evidence when reducing it to text would discard context needed later. Schema validity does not make payer information true. Different payer systems can disagree or update later, so an internal X12-shaped record should remain provisional until downstream evidence confirms or contradicts it.
This identifies a stable industry protocol as a common intermediate representation for agents acting across otherwise different interfaces. X12-shaped actions make fields, permissions, and lifecycle transitions mechanically checkable while retaining multimodal evidence for later judgment. It also separates protocol conformance from truth: valid records remain provisional when payer systems disagree or change.