The vision is one canonical model, one audit trail, one role system, every stage AI-assisted with human review gates. Two modules are live today. The rest are on the build path.
Author M11-aligned protocols section by section with a Claude-powered co-pilot. Every Code field binds to CDISC CT. The CORE rules engine runs as Layer 3 of the pre-export compliance gate — exports block on ERROR-severity findings unless explicitly overridden with an audited reason.
Read a USDM-based study definition (or a protocol PDF, or a USDM-JSON export) and generate a CDASH-aligned casebook. Forms, fields, codelists, edit checks, derivations — derived from the protocol's structured content. Deploy to Medidata Rave or Veeva Vault.
Each roadmap module carries its target scope here so pilots and prospective customers can plan ahead. We don't commit dates publicly — we commit milestones. Early-access pilots can include scoped previews of in-development modules.
Unified pipeline for any EDC + any vendor. Rave (ODM.XML or Clinical Views), Veeva Vault, central labs (SAS7BDAT / XPT / CSV), imaging, RTSM, ePRO, devices. Every row carries source-system OID for idempotent upserts.
Three-level rule engine — syntactic + format, logical edit checks within and across forms, cross-source consistency. Every finding ships with rule ID, evaluated logic, observed values, severity. Findings route to Query Management.
One lifecycle across validation findings, reconciliation discrepancies, coding ambiguities, manual reviews. Status states detected → drafted → issued → answered → closed. SLA tracking, aging dashboards, escalation rules.
Verbatim AEs, medical history, concomitant medications. Auto-coder proposes LLT / PT, ATC, or trade-name matches with confidence scores. High-confidence auto-codes; below threshold routes to coder workbench.
SAE sub-agent matches safety-DB cases against EDC AEs / SAEs. External sub-agent reconciles vendor data (labs, ECG, imaging, RTSM, ePRO, devices) vs. canonical EDC. Discrepancies route to queries or vendor tickets.
Mapping specifications from canonical data to every SDTM domain. Conformance checks, Define.xml + annotated CRF + reviewer guide. ADaM scaffolds — ADSL, ADAE, ADLB, ADTTE — with BDS structure for custom analyses.
Role-aware dashboards from portfolio → study → subject. Open / aged queries by type and site. Reconciliation completeness. Coding completion. CRF timeliness. Lock-readiness gauges. Risk flags on site / subject / domain anomalies.
Per-subject longitudinal profiles — demographics, visits, exposure, AEs, labs, vitals, ePRO. Risk flags surface anomalies. Treatment-arm fields hidden from blinded roles at the repository layer (not just the UI).
Append-only audit across every module. Tamper-evident storage. Searchable by study, subject, user, record type, event type, date range. Inspection-ready PDF + machine-readable JSON export. AI actor distinguishable from human actor.
Eleven predefined roles, extendable. Fine-grained permissions per module + per data domain + per blinding scope. MFA / OTP. Separation of duties. In-app guidance, walkthroughs, citation-verified AI Guide.
Thirty-minute working session — not a deck. Live USDM authoring against your protocol; a CDASH study build from the same source. Early-access pilots welcome.