AI-Powered · In Production-Ready Development · US Healthcare Operator (Under NDA)
A 3-module AI-augmented ecosystem — Super Admin governance layer, multi-tenant Agency Dashboard, and field-grade Caregiver Mobile App with biometric EVV. Document intelligence (OCR/NLP), voice-to-text caregiver notes, predictive scheduling, risk-flag detection, and anomaly detection are embedded across the stack — not bolted on. Built for HIPAA, 21st Century Cures Act EVV, CMS Medicaid/Medicare, EDI 837/835, and SOC 2–aligned controls.
A US healthcare operator approached us to build the next generation of adult home care agency management software — a multi-tenant SaaS that home care agencies could subscribe to instead of running fragmented legacy tools. The problem in adult home care is structural: agencies juggle paper intake forms, separate scheduling tools, third-party EVV apps, manual billing pipelines, and unconnected payroll exports. Compliance breaks down at every handoff. Caregivers can’t verify visits without 4-step workarounds. Billing claims get rejected for missing EVV data. Audit prep takes weeks.
The client needed a unified platform where one agency subscription replaces five disconnected tools — with HIPAA-grade security, federally-mandated EVV compliance, EDI 837/835 billing, and a mobile caregiver app that survives an audit. Built once, sold many times to home care agencies across multiple US states. And critically, they needed AI engineered into the platform from day one — not as a marketing layer over a legacy product, but as a real capability set covering document intelligence, voice-to-text caregiver workflows, predictive scheduling, and risk-pattern detection. The goal: ship a platform that’s structurally ahead of the legacy home-care SaaS competition on launch day.
We are architecting and building a complete 3-module SaaS ecosystem covering the full lifecycle of an adult home care agency.
HIPAA, 21st Century Cures Act EVV, and CMS Medicaid/Medicare requirements drove the data model, the deployment topology, and the audit-log design from week one — not bolted on post-build.
Each agency tenant gets a dedicated database schema within a single database instance. Tenant-aware routing and query scoping enforce zero cross-tenant data access at both application and database layers.
The Super Admin platform handles tenant provisioning, subscriptions, and platform configuration — but never accesses, processes, or stores Protected Health Information. PHI stays exclusively inside tenant schemas, by architecture not by policy.
Biometric login (Face ID/Touch ID), face-detection at clock-in/out for anti-fraud, GPS-verified visit location, device fingerprinting, and JWT-based session security on a Flutter app engineered for unreliable field network conditions.
EDI 837 (claims) and EDI 835 (remittance) integrations designed in from the data model, not retrofitted. Payer-specific billing rule engine. Payroll export pipelines for OT, holiday, and weekend rule handling.
AI · DIFFERENTIATION
Legacy home-care SaaS competitors typically retrofit AI as a marketing layer over a 10-year-old product. We’re engineering AI into this platform from the data model up — six concrete capabilities that reduce caregiver documentation burden, surface clinical risk earlier, automate the highest-error data-entry steps, and remove operational friction across the agency.
Extracts structured data from insurance cards, prior authorizations, medical intake forms, and uploaded clinical documents automatically. Eliminates one of the highest-error data-capture points in the agency workflow and reduces intake time from hours to minutes.
Caregivers dictate visit notes in the mobile app; AI transcribes in real time, applies grammar cleanup, and structures the output for the clinical record. Designed so caregivers can document while still in the client’s home — not three hours later from memory.
AI analyses visit notes, ADL completion patterns, and visit-frequency signals to flag clinical risk indicators — fall-risk changes, medication non-adherence patterns, hospitalization likelihood — and routes them to RN supervisors before they become incidents.
AI suggests optimal caregiver-client matches based on geographic proximity, skill match, language preferences, historical care quality, and availability — instead of coordinators solving the matching problem manually on a spreadsheet for every shift.
EVV streams are continuously analysed for anomalies: clock-in patterns that don’t match GPS data, unusual visit-duration spikes, suspiciously identical visit notes across clients — surfacing potential fraud or data quality issues before payer audits do.
Long-form caregiver notes are summarised for RN review; documentation quality is scored against compliance standards; flagged notes are routed for caregiver coaching. Improves the audit-readiness of the clinical record over time without manual chart review.
Schema-per-agency partitioning enforced at the database layer with middleware validation that runs before any business logic executes — making cross-tenant data exposure architecturally impossible, not just policy-prevented.
Combined GPS, biometric, and face-detection verification at each caregiver clock-in/out. Designed to satisfy state-level EVV adopter/aggregator requirements while remaining usable in the field.
Caregivers visit clients in homes with unreliable cellular coverage. Mobile app caches visit schedules and client details for offline read-only access, with EVV actions queued until reconnection. Sandboxed local storage, encrypted at rest.
The platform is in production-ready development. The numbers below describe architectural ambition and engineering scope — not historical production metrics.
Super Admin governance layer, multi-tenant Agency Dashboard, and Caregiver Mobile App — built to work as one product, sold as one subscription.
HIPAA Privacy & Security Rules, 21st Century Cures Act EVV, CMS Medicaid/Medicare, EDI 837/835, NIST Cybersecurity Framework, HL7, plus state-specific home care regulations.
Platform Owner, Super Admin (Operations / Billing / Compliance), Support Admin, Agency Admin, Coordinator, RN/Clinical Supervisor, Billing Specialist, Payroll Specialist, Caregiver — each with granular RBAC.
GPS + biometric + face detection at each clock-in/out, capturing caregiver identity, client identity, location, timestamps, services delivered, and ADL completion for full EVV mandate compliance.
Engagement Status
This platform is in production-ready development for a US healthcare operator under NDA. Architecture, compliance design, and module scope reflect engineering decisions made and code shipped to staging. Live production metrics will be added once the platform launches.
If you’re evaluating engineering partners for a regulated healthcare SaaS, three things matter more than any vendor’s sales pitch:
Tell us your clinical workflow. We’ll tell you the HIPAA architecture we’d use on day one.
Direct line, founder-to-founder +977 9851038796 — Roshan Subedi, Founder & MD · 48-hour reply, engineer-only
