-
Charting writes itself.
-
QA catches what gets missed.
-
Claims fly clean.
The back office for home-health and hospice. — more to come
It's late.
The system is still working.
AI-native infrastructure for home health.
Real-time clinical NLP.
Autonomous QA pipeline.
End-to-end revenue cycle.
Charting writes itself.
Your clinicians talk through a visit the way they’d describe it to a colleague — and the SOAP note is waiting when they sit down. No templates to memorize. No drop-downs to hunt for. No catching up at midnight.
Built for OASIS-E2 from day one, not retrofitted onto a generic scribe.
QA catches what gets missed.
A veteran nurse knows that “patient is independent in bathing” can’t sit next to “moderate assist required” three lines down. With a growing team, those inconsistencies slip through — and your QA staff spend their days reading every note to catch them.
Vesper reads every note the moment it’s signed. Narrative-code mismatches, inconsistent functional scoring, OASIS items that contradict the plan of care — flagged before a coder ever sees the chart. Your QA team stops line-editing and starts managing exceptions.
- OASIS-E2 consistency checks building
- Narrative-code alignment building
- Functional scoring validation designing
- Real-time flag-on-sign designing
Claims fly clean.
The chart that already passed QA writes its own bill. Primary diagnosis maps to the right clinical group. Functional scores land in the correct PDGM tier. Modifiers match. Comorbidities are captured, not guessed at.
Your billing team stops chasing clarifications. Denials drop. Revenue cycles that used to stretch weeks compress to days.
- PRINCIPLE
One source of truth from visit to claim.
- PRINCIPLE
The chart checks itself — humans handle exceptions.
- PRINCIPLE
Audit-ready by default, not on demand.
Real-time clinical NLP.
Live audio stream → speaker diarization → domain-tuned transcription → structured SOAP extraction. The nurse talks through a visit; the system parses it into clinical documentation that downstream OASIS assessments can rely on.
Not a generic whisper wrapper. The extraction pipeline knows which clinical details matter for OASIS-E2 — functional status, wound measurements, medication changes — so nothing gets lost between the visit and the assessment.
Autonomous QA pipeline.
Every OASIS assessment triggers a validation graph against the accumulated visit notes: narrative ↔ OASIS response consistency, functional scoring alignment across SOC/ROC/discharge, diagnosis-code specificity — no “unspecified pain” when the notes describe CHF with dyspnea.
The system that captured the visit documentation is also the system that checks the assessment — same context window, no information loss across the handoff. Flags route to human QA only when confidence drops below threshold.
- Cross-document consistency engine building
- PDGM clinical grouping validation building
- Contradiction detection (OASIS ↔ narrative ↔ POC) designing
- Confidence-gated human escalation designing
End-to-end revenue cycle.
Notes → OASIS → code → claim, one pipeline, zero re-keying. Primary diagnosis routes to one of 12 PDGM clinical groups. Functional impairment scores feed the three-tier adjustment. Comorbidity capture is systematic, not incidental.
432 case-mix groups. Two 30-day payment periods per certification. The model that understands the clinical picture also understands the reimbursement implications — because they’re the same data.
- PRINCIPLE
One model, one context window, visit to claim.
- PRINCIPLE
Every downstream artifact is a projection of the source note.
- PRINCIPLE
Audit surface is the pipeline itself — no reconciliation layer.