Claim Health

Models do the reading and drafting, and a workflow platform sits around them. Claim Health's AI reads referral packets arriving by fax, email and portal against the agency's own clinical criteria, citing findings back to the source, and drafts payer appeals from remittance detail.

Much of the rest is workflow the company runs for the agency: eligibility rechecks on every active patient, validation rules on assessment fields, authorization tracking against authorized hours, payment posting and reconciliation, and dashboards showing referrals, admissions and where care is stalled. Take the models away and an operations platform with rules, integrations and reporting remains, so the models are the engine of the intake and appeal modules rather than the whole product.

Completeness promises about the product's output sit beside a sensible division of work, and the promises are the problem. Claim Health's home page says it rechecks eligibility on every active patient 'so a quiet lapse never interrupts care', that Authorization Autopilot automates every authorization from submission to renewal, 'eliminating follow-ups and missed renewals', and that its workflows 'surface every dollar owed'.

Taken at their word, these promises remove the reason to keep the manual checks that catch what the system misses. The structure behind the product is reasonable: routine work is automated, exceptions go to the agency's billers, intake findings are cited back to the source and gaps flagged 'before anyone decides', and appeals are drafted with suggested actions for staff. The company does not say what counts as an exception, which actions are taken without a person, or how often something classed as routine goes wrong.

The approach is shown in the product, and nothing technical is named. Claim Health's home page describes AI that reads each referral packet against the agency's clinical criteria with findings cited back to the source, 'automated digital workers' for portal work, and drafted appeals built from remittance detail with suggested next actions. Assessment checks are shown as explicit validation rules on OASIS fields. No model, architecture, provider, version or accuracy figure is published, and nothing separates what a trained model does from what fixed rules and payer connections do.

Nothing establishes who else sits between a patient record and an answer. Claim Health's home page, privacy policy and terms name no model provider, hosting provider or subprocessor, and publish no retention or training terms for data that reaches any of them. The chain is likely to include a fax carrier for inbound referrals, the language model behind intake reading and appeal drafting, and the payer portals and clearinghouses it connects to.

Named customers and unsourced percentages, with no method behind either. The home page quotes executives at Home Care RN, Home Helpers and Ascend Health on faster, more disciplined revenue operations, without figures. It also reports 80 percent (manual hours), 15 percent (approved claims) and a 3 percent EBITDA lift under the heading 'Real results your team can feel in weeks', with no customer, baseline, period or method attached, and the case studies page lists no studies. The January 2026 funding announcement reports company revenue growth of 30 times in under a year, which measures adoption rather than benefit.

General assurances are published, and the questions an AI tool raises are not answered. The home page promises privacy first and HIPAA compliance, while the privacy policy covers only the website. Nothing states how long extracted referral packets, eligibility responses or claim data are kept, whether one agency's documents train models used for others, or whether anything is de-identified.

The material is broad: faxed and emailed referrals carry narrative about a patient's condition and care needs, not just claim fields, and post-acute agencies compete for referrals from the same hospitals. Safety engineering, meaning guardrails, hallucination mitigation and how a safety event is handled, is not described, although the intake tool reads clinical documents and the appeals tool drafts letters to payers.

HIPAA compliance is claimed and nothing that reaches the service stands behind it. The home page carries a 'HIPAA Compliant' badge. The privacy policy (November 2025) covers visitors to claimhealth.com only, and the terms of service cover the website only. No business associate agreement, statement of business associate status or service privacy document is published.

The platform reads each agency's record system and processes referrals and claims on its behalf, so it handles protected health information throughout. Referrals also arrive from hospitals and physician practices that are not Claim Health's customers.

Claim Health presents a SOC 2 attestation as its own, with nothing behind the badge. The home page's security and compliance strip reads 'HIPAA Compliant', 'SOC II Certified' and 'Privacy First'. No report type, period, auditor or scope is given, and no security page or trust center is linked from the site. These are small agencies with limited vendor review capacity, and the platform holds referral packets, eligibility data and claims from fax, email and portal intake.

No device claim is made, and the products are administrative. Claim Health checks eligibility, tracks authorizations, validates assessment fields, posts payments and drafts appeals. Its intake tool reads referral packets against the agency's own clinical admission criteria, cites findings back to the source document, and flags gaps before a person decides whether to admit; it does not recommend care.

The rules that matter here are Medicare's: home health and hospice face targeted probe and educate reviews, program integrity audits and pre-claim review in several states, and payment in home health is driven by the OASIS assessment, so the assessment is effectively the claim. The home page shows automated validation rules on OASIS fields, and the agency carries False Claims Act exposure for what is submitted.

Nothing is published on how model behavior is governed or tested. Claim Health's home page, privacy policy and terms carry no governance framework, model evaluation, monitoring description or bias review. The exposure is specific: the intake tool reads each referral against the agency's clinical criteria and flags gaps before an admission decision, and coverage checks flag patients whose payment is at risk.

The same signals that help an agency accept more patients can help it decline those with complicated coverage or incomplete records, who are not the patients who need care least. Nothing says whether flag rates or admission outcomes are monitored by patient group or payer.

Mechanisms let staff check what the system found, and nothing stands behind a wrong result. Intake findings are cited back to the source document and gaps are flagged before a person decides on admission; exceptions go to the agency's billers; appeals are drafted with suggested actions rather than sent unseen. No warranty, remedy, service credit tied to outcomes or measured error rate is published, and the website terms disclaim liability for use of the website only. For these agencies the costly errors are a missed eligibility lapse, a lapsed authorization and a claim submitted on a wrong assessment field.

Integration is described without a record system named. Claim Health says it monitors coverage, documentation and eligibility inside the agency's record system, validates OASIS assessment fields against conditional rules, gathers referrals from fax, email and portals, and works with payer portals and clearinghouses such as Availity for eligibility and claims.

The home page shows it checking records in 'your EMR', but neither the site nor the launch announcement names a home health, hospice or home care record system (Homecare Homebase, WellSky, Axxess, KanTime and MatrixCare are the usual candidates), an interface method or a production deployment on one. Clearinghouse and payer portal connections are not record system integration.

Nothing is published about where the system runs or where agency data rests. Claim Health's home page, privacy policy and terms name no hosting provider, region, tenancy model or residency commitment. Referral packets, eligibility responses, assessment data and claims all pass through it, and faxes arrive through a third-party fax carrier before they reach the platform.

Claim Health sells through demo requests to home based care agencies, and nothing about cost is shown ahead of that conversation. No price, unit of charge, module pricing or implementation fee is published, and nothing says whether fees are a subscription or a share of collections. For a product that works claims to cash, a percentage of collections would tie the vendor's income to how hard claims are pursued.

Coverage is named for home-based care, with validation behind part of it. Claim Health sells to home health, hospice, home care and other home-based care agencies, and its launch announcement adds nursing homes among post-acute providers. Named customers speak for it on the home page (Home Care RN, Home Helpers, a multi-state organization, and Ascend Health), but they give no figures, and the case studies page lists no studies.

Nothing states which payer mixes, states or agency sizes the product has been proven in. The products themselves cover intake and authorization. Smart Intake gathers referral packets from fax, email and portals and checks each against the agency's criteria, Coverage Assurance rechecks eligibility for every active patient, and Authorization Autopilot handles prior authorization from submission to renewal.