The rule in one sentence: when you connect an AI assistant, the data you ask for is shown inside that assistant under its provider's terms; Genaya does not share your data with the assistant provider on its own. Genaya answers the question you asked, limited to what you can see, and hands that answer to the assistant. What happens to it from there - inside ChatGPT or Claude - is governed by that assistant's own terms and by your own settings with it (for example, whether your chats there may be used for training), not by Genaya.
What leaves Genaya
The answer to the question you asked, and only that. Nothing is synced, exported or copied in the background.
Only what your role can already see in Genaya, in the one organization you picked - the same permissions, the same company scope, the same hidden fields.
Text a customer could have written (notes, messages, form answers, reviews, transcripts) comes back marked as customer data, shortened and cleaned, so it is treated as information and never as an instruction to the assistant.
What never leaves Genaya
Card, bank, tax, payroll and credential data, and the internal fields hidden from your role - none of it is available to any assistant tool.
Your Genaya password or sign-in codes. The assistant never sees them: you sign in on a Genaya page, and the assistant receives a connection token that expires after an hour and renews itself quietly. Genaya keeps only a fingerprint of that token, never the token.
A typed step-up value (for example a confirmation for a large amount or a changed phone number) - it is typed on the Genaya page and never passes through the assistant.
Settingschevron_rightRoles and Permissions
shield_personManager8 members
checkWhat the assistant may do
checkRead what you can see
checkAsk Genaya AI as you
Never from an assistant
Card, bank, tax, payroll data
Money, billing, roles, payouts
An assistant inherits your permissions exactly - what you cannot see in the app, it cannot see either.
What Genaya keeps
Genaya records each connection (the assistant, the organization, when it was created, when it was last used and with which tool) and one log line per call that names the tool, the assistant and the timing - never the question, the answer or the token. Every question also appears as a conversation in your Genaya AI history in the app, marked with its source, exactly like a chat you typed there. The Genaya AI answer itself is produced by Genaya's own model providers under the AI features section of the Privacy Policyarrow_outward, the same way an answer in the app is.
The healthcare note
warningConsumer ChatGPT and Claude are not covered by a business associate agreement; do not share protected health information through this connection. For healthcare organizations external assistants are off by default and stay off until an admin acknowledges this note on the Organization section of the page. The same note appears on every consent screen in a healthcare organization.
Your Genaya workspace, and any agreement you hold with Genaya, does not extend to the assistant. If you are a covered entity or a business associate, treat ChatGPT and Claude as outside systems: ask about schedules, counts and business figures, not about a patient's condition or treatment.
Organizations and linked organizations
A connection is bound to one organization and sees only that organization. Linked organizations - a link between two businesses - are not team memberships and are never included in a connection; to use Genaya in another organization you are a member of, connect again and pick it. If your role, your plan or the organization switch changes, the assistant's access changes with it within a minute, and a revoke lands on the very next call.