Artificial Intelligence Disclosure

How Sage uses AI, what happens to your content, what Sage remembers, and the limits you should know about.

Sage is the personal health intelligence inside Veyda. This page explains how Sage's AI works with your information, the commitments we make about it, and the limits you should keep in mind. It supplements our Privacy Policy and Medical Disclosure.

How Sage works

When you talk with Sage, your message and the relevant parts of your health context, such as recent metrics, labs, and history you have connected, are processed by third party AI service providers to generate a response. Sage can also use tools inside Veyda on your behalf, such as logging a food or pulling up a chart, and it will show you what it is doing.

To keep answers grounded, Sage draws on your own connected data and on vetted health evidence sources. When Sage consults specialized clinical evidence services to support an answer, your question is de-identified first: direct identifiers are stripped before anything is sent.

Our commitments

• Your content is not used to train the third party AI models that power Sage. Our agreements and configurations with these providers prohibit using your conversations, health context, memories, and other Sage content to train their models.
• Healthcare grade handling for member conversations. Our primary AI provider processes member conversations under a healthcare grade agreement with heightened confidentiality obligations and limited, time bound retention. A backup provider may step in if the primary is unavailable, under agreements that prohibit training on your content.
• No advertising, no sale. Your conversations and health information are never shared with advertisers and never sold.
• De-identification for specialized processing. Clinical evidence lookups are de-identified before they leave our systems.
• Quality and safety review. A limited number of authorized Veyda staff may review conversations in pseudonymized form, with identifiers masked and personal details redacted, to monitor quality and safety and to investigate problems you report.

What Sage remembers

Sage keeps memories so it can be genuinely personal: things you tell it to remember, and things it notices over time, like your preferences and how you like to communicate. You are in control:

• What Sage knows about you, in your profile, shows what Sage has picked up. You can correct any item or remove it.
• Memories Sage saves during conversations can be managed in Sage's memory settings.
• Deleting your account permanently deletes all memories.

Limits you should know

• Sage is informational and educational only. It does not provide medical advice, diagnosis, or treatment, and talking with Sage does not create a doctor patient relationship. See the Medical Disclosure.
• AI can be wrong. Sage works hard to be accurate and cites the evidence it draws on, but AI generated content can be incomplete, outdated, or incorrect. Verify important health decisions with a licensed clinician, and do not start, stop, or change a medication based on Sage alone.
• Sage does not make significant decisions about you. Sage offers guidance; it does not decide your access to services, insurance, employment, credit, or anything similar.
• Emergencies. Sage is not for emergencies. If you may be in danger, call 911. If you are in crisis, call or text 988.

Sage by text

If you activate Sage over text messaging, the same commitments apply, with one caveat: messages travel over standard messaging services operated by Apple, your carrier, and our messaging delivery provider, so transport is governed by those services. For sensitive topics, Sage will guide you back into the app. Details are in the Privacy Policy.

Questions

Email privacy@veyda.com, or reach us through support in the app.