The provider is Operator details will be published here, identified in the Terms; privacy requests go to support@ascendia.day. No AI interaction creates a professional or coaching relationship. Users must independently check every AI number, balance, reminder, schedule and suggested action before relying on it. The Privacy Policy explains what content may go to an external AI provider.
Effective 27 September 2026. Cognito and Ascendia's agents help draft plans, summarize notes, propose tasks and goals, answer questions and estimate food from an approved meal photo. The service may use your prompt, account context and prior information to create a response. The configured primary model route is Google Vertex AI/Gemini; Anthropic and OpenAI are optional providers. Optional Google Cloud speech and Voyage embeddings have separate processing paths.
AI responses are generated, not checked facts. They can be inaccurate, stale, incomplete, misleading, biased or offensive, can invent sources or numbers, and can misunderstand your request. Vision can mistake ingredients and cannot verify allergens. Cognito gives no medical advice, diagnosis or dosing; a health-related answer carries a short disclaimer below it. An AI suggestion is not mental-health, nutrition, training, legal, tax, investment or other professional advice. Check important output against primary records and consult a qualified person before acting. There is no guarantee of correctness, availability or a particular result.
You decide whether to accept a draft or recommendation. Some actions may be written by a tool after confirmation or according to autonomy settings you choose; inspect the proposed effect and the resulting receipt. We do not intend for an AI response by itself to decide access to an account, credit, employment, healthcare or another matter with legal or similarly significant effect. If such automated decision-making is introduced, it requires a separate notice, lawful basis and applicable review rights before use. GDPR Article 22.
You can review, correct or forget certain learned memories and pause future learning in Settings → AI Memory. Feedback about a response does not grant model-training consent. Provider retention and training depend on the active contract and settings.
To report a bad or unsafe output, use the response feedback or report action where offered, or write to support@ascendia.day with the conversation date and a description. Avoid emailing unnecessary health, financial or other sensitive details. For an immediate safety issue, use local emergency or professional help rather than waiting for review. The Privacy Policy explains which data may be sent to an AI service and how to exercise data rights.