SKYTIME AI is a complete home-health EMR where the AI drafts the notes, suggests the codes, scrubs the claims, and keeps unsigned orders moving — and your team reviews and approves every step.
Most EMRs bolt on a chatbot. SKYTIME's AI lives inside the chart: it stages its work at the exact field it belongs to, cites the source language for every suggestion, and nothing it writes can be signed until a human reviews it.
Upload the referral packet — demographics, payer, diagnoses, meds, and orders arrive pre-filled with source citations. Your coordinator confirms instead of retyping. Unmentioned means blank: it never invents data.
Record or dictate the visit and the note drafts itself — vitals pulled, skilled-care justification in the clinician's own words, and honest flags for what the dictation didn't cover.
M-item and ICD-10 suggestions with the exact note language cited on every one. Real coding rules built in — hypertension combos, Z79, diabetes linkages. Accept, edit, or clear each item; bulk-accept is view-gated.
Every claim checked before it leaves — PDGM groupability, cert and F2F consistency, dozens of denial-pattern checks. It flags and explains; your biller decides.
Unsigned orders age visibly, escalate on schedule, and show exactly which dollars they block. Portal e-signing, fax, and download paths — one pipeline, one signature chase.
While the office sleeps, SKYTIME sweeps every chart, stages suggestions inline, and flags risks — so staff arrive to a triaged morning, not a backlog.
Our AI doctrine, in one line: it guides, it never guarantees — and it never acts alone. Every AI feature carries a kill switch, an audit trail, and a promotion gate: no AI capability runs autonomously until it passes its own regression test set. When sources disagree, SKYTIME shows you the cited timeline — it never picks a side for you.
My Day computes each person's real worklist — patient, what's needed, why, and one verb. The dashboard counts and the worklists are the same query, so they can never disagree.
Every visit on a live map with clinician territories, day paths with mileage, and proximity outliers flagged. Type any address to see instantly who covers it before you accept the referral.
NOA and claim readiness computed from the chart itself — every blocked claim names its gate and links to the fix. SNIP-validated 837s, real 835 posting with denial routing, per-payer emission rules, and honest tracking for files you submit through a clearinghouse or portal.
SKYTIME continuously runs the checks a state surveyor or accreditor runs — against your live data, not a binder you assemble the week before. Every finding names the patient, the regulation, and the fix.
Permissions are modeled on the workflows agencies already know — each role lands on its own worklist, and the navigation only shows what that role can actually open.
My Day, own patients' charts and documentation, own schedule — scoped to their assignments, enforced server-side.
The referral pipeline with AI extraction, duplicate detection, eligibility checks, and document intake at every phase.
Calendars, the live map, staff availability, and chart summaries — enough to schedule, nothing clinical to wade through.
Coordination notes, the orders desk, document scanning, and med reconciliation proposals across all patients.
The billing center, eligibility hub, AR reports, and read-only chart evidence — verify everything, edit nothing clinical.
The coding queue, ICD picker, diagnosis worksheets, and the pick list. Coders code; QA approves.
The QA center, survey readiness, infection and incident logs, and full audit read across the agency.
All of it — plus the executive dashboard, setup, and the AI governance console with every kill switch.
Every status change is backed by an immutable event with an author — signatures, versions, and QA history preserved forever. Nothing is ever silently edited or deleted.
Granular permission grants, per-role session timeouts, TOTP two-factor, and full audit logging with export.
Identifiers masked on prints and claims views, sensitive-field reveal is itself a permission and an audit event, and print output carries exactly what the record supports — empty sections say "Not documented."
OASIS-E2 exports with pre-submission edit checks, PDGM grouping, SNIP-validated UB-04/837I, 835 remittance posting, and Cures-Act-complete EVV capture.
Every AI capability has a test set, a policy level, an audit trail, and a kill switch on one governance screen. Suggestions cite their sources; conflicting sources render as a timeline, never a guess.
Our data-migration process is rehearsed and documented: full backup, dry-run reconciliation you can read, count verification on every screen, and a rollback plan — before a single record moves.
SKYTIME AI is in staff testing with our first partner agency. Production onboarding — including HIPAA business associate agreements and payer enrollment support — opens with our early-access cohort.
Book a 30-minute walkthrough and we'll run your intake-to-claim workflow end to end. Or click around the live demo right now — it runs on synthetic data with every role available to try.
Open the live demo →Less painful than staying, and we've built for exactly this. SKYTIME's navigation and permissions are modeled on the workflows your staff already knows, migration follows a rehearsed runbook — backup, dry-run with a readable reconciliation, verification, rollback plan — and we handle the payer-enrollment paperwork alongside you. Your team's retraining burden is a design goal, not an afterthought.
The AI never signs anything and never acts without review. Every suggestion cites the source language it came from, editing a suggestion counts as reviewing it, bulk-accept is gated on actually viewing each item, and every capability has a kill switch your administrator controls. We treat "the AI guides, it never guarantees" as a hard product rule — you'll never see us promise that software prevents denials or survey findings.
Early-access pricing is set per agency with our first cohort — one conversation and you'll have a number. No per-module nickel-and-diming: the AI features are the product, not add-ons.
SKYTIME AI is built by CareOS AI LLC, an Illinois company, hands-on inside a Medicare-certified home health agency — every feature ships because a real coordinator, biller, nurse, or QA reviewer needed it, and their feedback lands in the product weekly.