Study overview
This descriptive methodological study documents Salmona, Arkangel AI's rule-grounded medical billing review architecture for Colombian healthcare reimbursement. It draws on technical documentation and the applicable Colombian regulatory framework.
Version 1 was posted on Research Square on June 30, 2026. It is a preprint under review, not yet peer reviewed by a journal. No patient data were collected and no empirical validation was conducted.
Three traceable audit layers
- Administrative: 27 DAMA-UK rules.
- Clinical necessity: 29 PERT-CLIN rules examining evidence supporting billed care.
- Financial and contractual: 42 FIN-CTR rules.
Each rule returns pass when compliance is supported, fail when a violation is supported, or n/a when the required information is absent. Missing documentation does not automatically become a billing objection.
Human governance
A consolidation layer reconciles findings per invoice item, assigns objection causals under Technical Annex 6 of Resolution 3047 of 2008 and calculates objected amounts under a financial integrity invariant. A human auditor reviews findings, requests corrections through comments and explicitly initiates report delivery to the provider.
Limitations and disclosure
This is an architectural description, not evidence of accuracy, savings, reduced billing objections or clinical impact. Validation with real claims remains future work. The authors also describe potential provider-side pre-submission validation and structured responses to objections. All authors declare employment at Arkangel AI, which developed Salmona. The manuscript is licensed under CC BY 4.0.
Original source
Read the Research Square preprint · DOI: 10.21203/rs.3.rs-9819774/v1




