Evidence before claims.
What we publish, what we share on verified corporate email, and what requires an NDA and a call. We do not claim to have solved primary care, we do not claim clinical accuracy from the post-market report, and we avoid "world's first". The open questions are listed below the evidence.
Public
Clinical white paper, February 2026
Overcoming the Cognitive and Sensory Deficit in Modern Medicine, Dr. Syed Sabahat Azim, MBBS, former IAS. The clinical rationale for physical AI: the snapshot fallacy, the safety gate, longitudinal multimodal modelling, and evidence-based prescribing at scale. Freely downloadable and citable.
Working monograph, SSRN preprint
Beyond Primary, Secondary, and Tertiary: A 2×2 Healthcare Framework. A practice-derived replacement for the tier classification: two axes, acuity and intervention dependency, four operationally coherent quadrants. Preprint; not peer-reviewed.
Post-market clinical experience, summary
Report reference ZoyeMed3.0-PMCF-2026-01, dated 6 March 2026. 18,224 consultations, Mexico institutional programme, March 2025 – February 2026, nurse-assisted workflow with remote specialist supervision. Reported: high operational uptime, strong user satisfaction, zero serious adverse device effects, zero device incidents involving patient harm; client-attributable complaints (operator training, local infrastructure) exceeded device-attributable complaints. The report demonstrates safety and operational feasibility, not clinical accuracy.
Lineage citation, 2015
LitmusDX, the predecessor clinical decision support system deployed across the Glocal Healthcare network, was one of thirteen systems documented in Shahsavarani and Abadi's 2015 state-of-the-art review of clinical decision support systems. Evidence of lineage recognition, not of current-generation validation.
Verified corporate email
Technical white paper, April 2026: The Category Nobody Else Is Building, a thirteen-platform, seven-axis comparison and the PMCF summary. Interactive 3D terminal model: sensor positioning, operator workflow, footprint and ports. Delivered automatically after one-time-password confirmation; free webmail addresses are not accepted.
NDA and a call
Full post-market report with patient-level safety, operational performance and AI–clinician concordance data. IP technical disclosure covering multimodal fusion, the GRU-family and state-space sequence cores, sufficiency gating and human-in-the-loop orchestration (PCT application filed Q1 2026, with a patent of addition). Financial model and commercial terms.
Open questions we name ourselves
Large-scale deployment outside Mexico. Prospective, peer-reviewed clinical-accuracy studies of the current longitudinal model. Multi-region closed-loop outcome integration. Federated learning at multi-country scale. Long-horizon model-drift analysis. All treated as work in progress, not as settled.
What we do not claim
- That the 2015 citation validates current performance
- That the post-market report demonstrates diagnostic accuracy
- That our deployment numbers exceed peers' (several platforms have served far more patients)
- Any capability not confirmed in published documentation
- "World's first" anything
Research hub: writing
Essays and papers from the team, including the sensory-deficit argument for physical AI and the four-quadrant framework applied to maternal and child health. Writing