Pharmacy operations exhaust generated by running dispensing work end to end instead of recording it: prior-authorisation packages assembled from claim rejections with an approval-probability score per case and payer follow-up tracked on a knowledge graph; benefit-verification and coverage-check events; outbound/inbound payer and patient CALL RECORDINGS with their structured outcomes; and the workflow layer the vendor says 'learns from your pharmacists' — i.e. per-queue, per-handoff decision traces of how pharmacies actually resolve each case.
Equities in US. Related tickers CVS, WBA, COR, PNTM, UNH, ELV, HIMS, LLY.
Multimodal: audio (payor/patient voice calls, IVR and SMS transcripts), tabular case/claim/coverage records, and document text (prior-authorisation packages, appeal and clinical-review artefacts)
Large for a seed-stage firm: millions of minutes of captured workflow and call time per year, tens of thousands of PA and benefit-verification cases, and per-case artefacts (package documents, probability scores, call audio, correction records) — audio being the scarcest component
Case-level resolution labels — which PA package content led to approval versus rejection, with the approval-probability score recorded before the outcome (a pre/outcome pair usable as a reward signal); 'handled without human intervention' flags on calls, which label each call segment as successfully resolved or escalated; OCR-to-verified-field correction pairs on handwritten prescriptions and faxes (what the AI typed versus what the pharmacist fixed); payer/queue routing decisions; task-duration labels (minutes saved per case) that make this a productivity ground-truth set; prescription-intent labels from intake.
CareLink (by Asepha) — Pharmacy operations exhaust generated by running dispensing work end to end instead of recording it: prior-authorisation packages assembled from claim rejections with an approval-probability score per case and payer follow-up tracked on a knowledge graph; benefit-verification and coverage-check events; outbound/inbound payer and patient CALL RECORDINGS with their structured outcomes; and the workflow layer the vendor says 'learns from your pharmacists' — i.e. per-queue, per-handoff decision traces of how pharmacies actually resolve each case.
Not yet on file: Sample · Data dictionary · Data frequency · Reporting lag · Point-in-time · License