Last updated: 2026-09-30
RxAiAnalytics LLC ("RxAiAnalytics," "we," "us," or "our") provides a cloud-based software service (the "Service") that analyzes pharmacy-claims data for health plans, plan sponsors, brokers, and pharmacy benefit managers (PBMs). This Privacy Policy describes how we handle information in connection with our website at rxaianalytics.com and the Service.
RxAiAnalytics is a business-to-business (B2B) service. It is not a consumer product and is not directed to the general public. Two very different kinds of information are involved in our business, and this policy keeps them separate throughout: (1) the pharmacy-claims Protected Health Information (PHI) that our customers upload, which is governed by HIPAA and by the Business Associate Agreement (BAA) between RxAiAnalytics and each customer; and (2) the account, business-contact, and technical data we collect directly to operate our website and the Service, which is what this Privacy Policy primarily governs. Please read Section 1 for how these two categories differ and where each is governed.
RxAiAnalytics LLC operates the Service that lets authorized users from approved customer organizations (health plans, plan sponsors, brokers, and PBMs) sign in and upload pharmacy-claims data for analysis — including savings and biosimilar detection, anomaly detection, contract analysis, and a natural-language question-and-answer assistant. Access is restricted to approved organizations (via an email-domain allow-list) or to users invited by an administrator. The Service is not directed to children.
This Privacy Policy covers the two categories of information described below and explains, at a high level, how PHI is protected. It does not replace or override the BAA.
Individuals who are plan members and want to exercise rights over their PHI (such as access, amendment, or an accounting of disclosures) should contact their health plan — the covered entity — as explained in Section 9. As a Business Associate, RxAiAnalytics cannot independently fulfill individual HIPAA rights, but we assist the covered entity in doing so.
Account and business-contact data. When an organization is approved and its users are provisioned or invited, we collect information needed to create and manage accounts, such as name, work email address, the company or organization the user represents, role, and authentication data managed through our identity provider, Amazon Cognito (including multi-factor authentication enrollment).
Technical and usage data. To operate and secure the Service, we process functional and technical data such as authentication and session tokens stored in the browser, IP address, device and browser information, request identifiers, and application and security logs that record access to and activity within the Service.
Support communications. If you contact us for support or otherwise correspond with us, we collect the information you provide and records of that communication.
Billing and purchase data. If your organization purchases a self-serve plan, checkout is handled by our payment processor, Stripe, on Stripe-hosted pages. We receive and retain transaction records — the plan purchased, amount, date, payment status, and the billing contact and address associated with the purchase — but full payment-card numbers are collected by Stripe directly and never touch our systems. Stripe's handling of payment information is described in Stripe's own privacy policy.
We use account, technical, and support information to:
When we work to improve the Service, any use of customer data for that purpose is limited to aggregated data with no member identifiers, such as the aggregate drug-to-drug transition counts described below, and is carried out consistent with the BAA. Model training: Customer content is contractually excluded from training Anthropic's models, and RxAiAnalytics does not provide customer data to any third party to train models. Inside the platform (AWS), RxAiAnalytics runs its own statistical models on your claims: spend forecasts and member risk flags built from your own uploads, and — unless you turn it off for a client or an upload — cross-client learning, which stores only aggregate drug-to-drug transition counts with no member identifiers and uses them to calibrate the drug-progression risk flags of every participating upload. Both can be turned off per client or per upload; turning them off removes that upload's results and its contribution.
PHI is handled under the terms of the BAA and HIPAA. At a high level, the safeguards we apply to PHI include:
Requests by individuals to exercise HIPAA rights over their PHI are directed to the covered entity, as described in Section 9.
Several features of the Service are AI-assisted, including the natural-language question-and-answer assistant, contract analysis, and anomaly detection. These features are powered by Anthropic's Claude, which acts as an AI subprocessor.
What is transmitted differs by pathway, and each is enforced in our code. Analytics requests — the assistant, the claims context supplied to contract analysis, and the plan-audit narrative — carry aggregated data only (apart from the selected-contract analysis and the optional member-level detail described below): totals, spend, claim counts, distinct-member counts, and cost and utilization measures, grouped by drug and drug attributes, condition, tier, pharmacy channel and type, network, days supply, plan (as “Plan 1”, “Plan 2” and so on), prescriber specialty, age band, gender and month. Where the assistant's detailed breakdowns are enabled, each breakdown group is reported with its exact counts, however few members it covers; the file summary sent with each question lists the top drugs and conditions with their claim and member counts. High-cost members are referenced by positional rank rather than identity, and prescribers only by specialty, never by NPI or name. No member identifier and no tokenized member ID is included, and your organization's name, like client, plan and contract-file names, is replaced with a placeholder before transmission and substituted back within our environment. High-cost claimants listed in carrier reports are likewise referred to by position rather than by name or printed label. When you select contracts in an assistant conversation, their stored analysis is also transmitted, with names replaced by placeholders; because it derives from a document you uploaded, it is a limited disclosure of PHI under our Business Associate Addendum. When you continue a saved assistant conversation, up to four of its earlier questions and answers are sent again as context. Member-level detail is off by default. It is available only in an assistant conversation, and only when your organization has allowed it for the upload and, where the upload is filed under a client, for that client, and the user chose it when starting the conversation; each of these is re-checked on every question. In such a conversation the assistant may retrieve individual pharmacy claim rows for members of that one upload: the masked member ID the Service displays on its member screens, exact fill date, drug name, NDC, GPI, therapeutic class, brand or generic, quantity, days supply, cost fields other than any your organization masks for a field broker, pharmacy channel, age band and gender, for at most 10 members and 200 rows per lookup, 40 rows per member, and 3 lookups per question. This is a limited disclosure of PHI under our Business Associate Addendum, not de-identified data. Those rows never include member names, the member or subscriber IDs in your files, Social Security numbers, dates of birth, addresses or contact details. Rows in restricted drug classes are withheld (and the assistant will not build a member list defined by one of those classes), web search is not used in that conversation, each member whose rows are transmitted is recorded in our accounting of disclosures, and member-level conversation content is deleted after 90 days. Schema detection works differently and we describe it precisely: to map the columns of an uploaded claims file the Service transmits a five-row sample in which a fail-closed classifier masks the values of every column it does not recognize as clinical, financial or operational data, while real dates of service and provider NPIs are transmitted because a format can only be detected from real values. Dates relating to an individual are identifiers under 45 CFR 164.514(b)(2)(i)(C), so we treat schema detection as an identifier-masked limited disclosure of PHI under our Business Associate Addendum rather than as de-identified data. Documents you upload for analysis are transmitted so they can be analyzed, and may contain your organization's name in their body text. Customer content is contractually excluded from training Anthropic's models, and we do not provide customer data to any third party to train models; our own in-platform statistical models, and how to turn them off, are described under How We Use Information.
Our AI subprocessor is engaged under a Business Associate Addendum accepted through a HIPAA-configured API organization, and customer content is contractually excluded from training its models. Text you type into the assistant and documents you upload are pattern-scrubbed for direct identifiers before transmission; because this is free text, scrubbing is a safeguard rather than a guarantee, and submitting PHI in free text is prohibited by our Terms. In a member-level conversation you may type the masked member IDs and dates the Service displays, which are transmitted as described above; names, raw member IDs, Social Security numbers and dates of birth remain prohibited. Outside member-level conversations, the assistant can also use the provider's web search, an in-scope covered service under that Addendum, where query content may be shared with third-party search providers. Every disclosure of PHI-derived context to the AI subprocessor is recorded in our HIPAA accounting of disclosures. AI-assisted outputs are intended to support decision-making and may contain errors; they are not a substitute for professional judgment.
We use only functional and technical browser storage that is necessary to operate the Service — for example, authentication and session tokens that keep you signed in. We do not use third-party advertising trackers or third-party analytics trackers.
We do not sell personal data, and we do not engage in cross-context behavioral advertising.
We do not sell personal data. We share information only as needed to operate the Service and as described here. We engage a limited set of subprocessors, all located in the United States:
Reference-data sources (no PHI). To support its analyses, the Service also queries external reference-data sources on a read-only basis with no member context: GoodRx (retail drug pricing), CMS/Medicaid NADAC (public dataset), and openFDA/FDA (public drug data). These sources receive no PHI and are not customer-data subprocessors.
We notify affected customers before adding a new subprocessor that will process PHI. We may also disclose information when required to do so by law, legal process, or a valid governmental request, or to protect the rights, safety, and security of our users, the public, or RxAiAnalytics — in each case consistent with the BAA for any PHI.
We retain account, business-contact, and technical data for as long as needed to provide the Service, maintain security, and meet our legal and contractual obligations, after which it is disposed of securely in accordance with our written data retention and disposal policy.
PHI retention and secure disposal are governed by that written policy and by the BAA. At the end of a customer contract, PHI is returned to the customer or destroyed in accordance with the BAA. The assistant saves each conversation — its questions, its answers and the evidence behind each answer — with the upload it relates to. A saved conversation can be opened only by the user who created it; that user can delete it at any time, and it is deleted when the upload is deleted, when that user's account is deleted, and at the end of the customer contract. Questions and answers that include member-level results are deleted automatically 90 days after they are created, and the individual claim rows the assistant read are not stored with the conversation. Machine-learning results for an upload are deleted, and its contribution to cross-client learning removed, when the upload is deleted or when either machine-learning setting is turned off for it. Deleted conversations and results remain in encrypted backups until those backups expire under our retention schedule.
Account and business-contact data. If you have an account, you may contact us to access, update, or request deletion of your account information, subject to our legal and contractual obligations and to verifying your identity. Depending on your jurisdiction, you may have additional rights regarding personal data; we will handle such requests in accordance with applicable law.
We maintain a written security and privacy program, overseen by a designated Security & Privacy Officer, and we apply reasonable and appropriate administrative, technical, and physical safeguards to protect information. These include multi-factor authentication, role-based access control and least-privilege access, encryption in transit and at rest, monitoring and alerting, incident response, and vendor risk management.
The Service is a B2B tool intended for authorized users from approved organizations. It is not directed to children, and we do not knowingly collect personal information from children under 13 (or under 16 where applicable). If you believe a child has provided us personal information, please contact us so we can take appropriate action.
The Service is hosted in the United States on Amazon Web Services in the us-east-1 region, and our subprocessors are located in the United States. The Service is intended for use by United States-based customer organizations. If you access the Service from outside the United States, your information will be processed in the United States.
We may update this Privacy Policy from time to time. When we do, we will revise the "Last updated" date above and post the updated policy at rxaianalytics.com/privacy. If we make material changes, we will provide notice as appropriate. Your continued use of the Service after an update takes effect means you have reviewed the current version.
Questions about this Privacy Policy or our privacy practices can be directed to our Security & Privacy Officer:
Security & Privacy Officer
RxAiAnalytics LLC
Email: privacy@rxaianalytics.com
Mail: RxAiAnalytics LLC, [Mailing/notice address — to be confirmed]