What AI actually automates in the home health billing cycle — and where Medicare coding, PDGM, and OASIS still require a specialist.
When home health agency owners ask about AI billing, they usually mean one of two things: AI that handles the operational workflow — documentation tracking, invoice drafting, unpaid claim follow-up — or AI that replaces the billing specialistwho handles Medicare coding, OASIS assessments, and PDGM calculations. These are very different asks, and confusing them is how agencies end up with billing software that doesn't do what they hoped.
The honest answer in 2026: AI handles the operational layer of home health billing reliably. It reads completed visit records, checks documentation before billing periods close, drafts invoices, tracks outstanding claims, and follows up on unpaid items — without someone having to monitor every step manually.
AI does not handle the clinical and compliance layer: OASIS scoring, ICD-10 diagnosis coding, PDGM payment grouping, or Medicare claims adjudication. Those require a licensed clinician and a billing specialist who knows Medicare home health payment rules. The two layers work together — but they are not the same layer.
The two terms are often used interchangeably, but the billing requirements are structurally different.
Non-medical services — companion care, personal hygiene, homemaking — billed primarily to private pay, Medicaid personal care waivers, long-term care insurance, and VA programs. Invoices are visit-based: hours delivered, service type, and rate per payer. The compliance layer is simpler — no Medicare coding, no OASIS, no clinical grouping. AI handles this billing layer well.
Skilled nursing, physical therapy, occupational therapy, and speech-language pathology delivered at home under a Medicare or Medicaid-certified home health agency. Billing uses the PDGM payment model — clinical groupings from ICD-10 codes, OASIS assessments, 30-day payment periods with LUPA thresholds, RAP and final claims. The compliance layer requires a billing specialist with Medicare home health expertise.
Both types of agencies share the same operational billing problem — ensuring documentation is complete before invoicing, tracking which claims have been paid, and following up on outstanding items. AI handles this shared operational layer in both settings. The Medicare compliance and clinical coding layer is unique to home health.
These are the operational billing tasks AI takes off the coordinator's plate — not clinical decisions, but the workflow layer that surrounds them.
AI monitors whether visit notes are complete and signed before the 30-day payment period closes. Missing or unsigned notes get flagged to the responsible clinician before they cause a billing hold — not after the claim is rejected or the LUPA threshold is missed.
AI tracks visits per 30-day period against the LUPA threshold for each patient and flags periods approaching the cutoff. Coordinators see which patients are at risk of a LUPA adjustment before the period closes — not when the final claim is processed.
When documentation is complete, AI reads the visit records and drafts the corresponding invoice at the correct rate for that client and payer — without a coordinator manually re-entering data that already exists in the visit record. Exceptions and unusual configurations are flagged for review before submission.
AI tracks claim aging from submission and surfaces past-due items for follow-up on schedule. For Medicare claims, it flags when a final claim has not been submitted within the required window, and escalates claims that age past the agency's payment expectations for coordinator review.
This is the honest version — the part of home health billing that AI software cannot do reliably and should not attempt.
PDGM payment groupings are driven by the primary ICD-10 diagnosis code assigned at the start of care. Selecting the correct primary diagnosis — and the correct etiology-manifestation sequencing for combination codes — requires a credentialed coder or a billing specialist with clinical coding expertise. AI cannot make this determination accurately or take responsibility for coding compliance.
OASIS completion and scoring requires a licensed RN or therapist making clinical judgments about the patient's functional status, living situation, and care needs. OASIS scores feed directly into PDGM payment calculations. No AI billing software completes or scores OASIS assessments — that is done by a clinician in the agency's EMR.
PDGM grouping, payment rate adjustments, LUPA calculations, and outlier payment determinations are performed by the agency's Medicare billing system or billing specialist. AI supports the PDGM billing cycle — tracking documentation, surfacing LUPA risk, monitoring claim submission windows — but does not perform the payment calculations themselves.
When a Medicare or Medicaid home health claim is denied, the resolution process — reviewing the denial reason, gathering clinical documentation, deciding whether to appeal, and preparing the appeal submission — requires a billing specialist who understands the specific denial code and the payer's reconsideration process. AI flags denials and surfaces the relevant documentation; the resolution is a human decision.
AI medical billing for home health is the use of AI software to automate the documentation-to-invoice workflow in a home health agency — reading completed visit records, checking for documentation gaps, drafting invoices, tracking unpaid claims, and sending billing follow-up messages. AI medical billing for home health handles the operational layer of the billing cycle: making sure invoices are built from complete documentation and that outstanding claims are followed up on schedule. It does not replace a billing specialist for Medicare coding, OASIS clinical assessment scoring, or PDGM payment methodology calculations — those require human expertise in home health billing compliance.
Home care billing covers private-duty personal care — non-medical services like companion care, personal hygiene assistance, and homemaking — typically billed to private pay clients, long-term care insurance, Medicaid personal care waiver programs, and VA programs. Invoices are visit-based: hours delivered, service type, and applicable rate. Home health billing covers Medicare- and Medicaid-certified home health agencies that provide skilled nursing, physical therapy, occupational therapy, and speech-language pathology services at home. Home health billing uses Medicare's PDGM payment model, requires OASIS clinical assessments, ICD-10 diagnosis coding, and RAP and final claim submissions through a Medicare-approved billing system. The operational billing workflow — documenting visits, tracking outstanding claims, following up on unpaid invoices — is similar in both settings. The compliance and clinical coding layer is unique to home health.
AI can handle the administrative and operational layer of Medicare home health billing — tracking visit documentation, drafting invoices from completed visit records, monitoring outstanding claim status, and following up on unpaid claims. AI cannot handle the clinical and compliance layer of Medicare home health billing: OASIS assessment scoring, ICD-10 clinical diagnosis coding, PDGM payment grouping calculations, or RAP and final claim submissions that require a Medicare-certified billing process. Medicare home health billing requires a billing specialist or a certified billing system built specifically for the Medicare home health prospective payment system — AI automates the workflow around that process, not the coding and compliance decisions inside it.
PDGM — the Patient-Driven Groupings Model — is Medicare's home health payment methodology, in effect since 2020. Under PDGM, each 30-day payment period is grouped based on the patient's admission source, timing, clinical grouping (derived from ICD-10 primary diagnosis codes), and functional and comorbidity adjustments. The PDGM grouping determines the base payment rate before the Low Utilization Payment Adjustment (LUPA) threshold is applied. PDGM payment calculations require clinical coding expertise and familiarity with Medicare home health payment rules — they cannot be handled by general AI billing software. AI tools support the PDGM billing cycle by ensuring visit documentation is complete, tracking whether a 30-day period has met the LUPA visit threshold, and following up on outstanding claims — but the PDGM grouping and coding decisions require a home health billing specialist.
In a home health agency, AI billing automation handles the operational layer of the billing cycle: (1) Checking that visit notes are complete and signed before the billing period closes — flagging missing or unsigned notes for the clinician or coordinator responsible. (2) Tracking whether 30-day payment periods have met the LUPA visit threshold, surfacing periods at risk of LUPA before the window closes. (3) Drafting invoices from completed visit records and routing them for review before submission. (4) Monitoring outstanding claims and surfacing past-due items for follow-up. (5) Sending billing follow-up messages for unpaid invoices. What AI does not automate: OASIS scoring, ICD-10 coding, PDGM grouping calculations, or Medicare claims adjudication decisions.
OASIS — the Outcome and Assessment Information Set — is the standardized clinical assessment tool used for Medicare home health patients. OASIS data is completed by a skilled clinician (RN or therapist) at the start of care, resumption of care, recertification, and discharge. OASIS scores drive the clinical grouping component of the PDGM payment calculation. OASIS completion is a clinical task requiring licensed clinician judgment — it is not an administrative billing task that AI automates. AI supports OASIS-dependent billing by flagging when an OASIS assessment is due or overdue, monitoring whether assessment data has been submitted before the billing window closes, and tracking the documentation status of OASIS-triggered billing events. The clinical scoring and submission is done by the clinician in the agency's EMR.
AI billing for home health agencies monitors the status of submitted claims by tracking their aging from submission date, surfacing invoices that have gone past the agency's expected payment window, and flagging claims for follow-up with payers. For Medicare home health claims, AI can monitor RAP and final claim submission status and alert the billing coordinator when a final claim has not been submitted within the required window. The AI tracks these as operational billing workflow items — the actual follow-up with Medicare Administrative Contractors (MACs) or other payers is handled by the billing coordinator, who has access to the payer's provider portal and understands the claim's clinical history.
Any AI system that reads visit records, OASIS data, patient demographics, or billing information for a home health agency is handling Protected Health Information (PHI) and must meet HIPAA's administrative, physical, and technical safeguard requirements. The vendor must sign a Business Associate Agreement (BAA) with the agency before any data is accessed. This requirement applies to all AI billing tools, all integrated scheduling and documentation platforms, and any AI messaging tool that touches patient information. For Medicare-certified home health agencies, HIPAA compliance is a baseline requirement — not a differentiator. Any vendor who cannot produce a BAA is not eligible to touch your agency's data.
The complete guide to AI billing for home care agencies — reading visit records, drafting invoices, tracking unpaid claims, and following up. Covers private-pay, Medicaid, and insurance billing workflows.
Billing accuracy in home health starts with complete, validated visit documentation. This guide covers how AI turns caregiver check-ins into structured visit records and surfaces documentation gaps before the billing window closes.
The broader picture: what AI handles across the full home care back office — hiring, scheduling, charting, and billing — and where human judgment still leads.
Medicare-certified home health agencies must maintain current caregiver credentials for compliance. This guide covers how AI monitors renewals and automates follow-up so a lapsed credential never reaches the schedule.
The complete honest breakdown — six workflows AI handles reliably in 2026, four it can't fully automate, and how to use the distinction to decide where to start.
Which workflows to hand off first and in what order — a sequencing guide for owners who want to automate without disrupting care or losing control of compliance.
Private beta · a handful of agencies
We're onboarding a handful of home care agencies now — get on the list.
Prefer to talk? Book a demo →