Shift handovers are the most dangerous moment in healthcare. Nurses and PSWs spend 15–30 minutes per handover delivering verbal reports — and critical patient information gets lost, forgotten, or miscommunicated. With 3 shifts per day, that's up to 1.5 hours of lost time per unit daily, plus significant patient safety risk.

ShiftHandover AI solves this instantly. Staff paste their raw, unstructured shift notes and AI converts them into a clean structured report in seconds — with colour-coded risk alerts, urgent task lists, ranked patient priorities, and a concise shift summary. One click exports a PDF for physical handoff.

Every handover is saved to a searchable database powered by pgvector embeddings. Charge nurses can search across all past shifts using natural language — "fall risk patients", "comfort care", "post-op complications" — and get semantically ranked results with similarity scores. The system also automatically surfaces similar past handovers on every report page, helping staff spot recurring patterns and at-risk patients across shifts.

Zero hospital system integration required — staff paste notes and go.

Business Impact

  • Saves 15–30 minutes per handover × 3 shifts/day = up to 1.5 hours saved per unit daily
  • Reduces miscommunication errors at the highest-risk care transition point
  • Creates a full searchable audit trail of every shift handover
  • Semantic vector search lets charge nurses identify recurring patient risks across shifts
  • Zero onboarding friction — no training required, no system integration needed

Built With

  • claude-api
  • jspdf
  • neon
  • nestjs
  • next.js
  • pgvector
  • postgresql
  • prisma
  • tailwind-css
  • voyage-ai
Share this project:

Updates