After AI consent, the signed-in assistant can use your saved name, self-reported ability and public favourites without treating them as verified skill.
Quick answer
Ask FASS can use minimal saved preferences to personalise recommendations after AI consent. Tell it what you want now: your current request takes precedence, and missing or outdated preferences should never be guessed.
After AI & Chatbot consent, the signed-in live assistant can automatically provide OpenAI with your saved account name, supported preferred activity, self-reported experience level and selected piste comfort. It can also use up to eight currently public favourite schools, roster instructors and resorts of each type. This helps it suggest relevant choices without asking you to repeat every preference.
Your current request comes first. For example, tell it “I usually ski, but this trip is for beginner snowboarding” and it should use that request rather than your saved skiing preference. Unknown values remain unknown. Numeric search-filter IDs must come from actual current taxonomy records, not guesses.
For learner accounts, the assistant may receive the number of completed bookings linked to the account and their recorded scheduled hours. These are booking records, not verified personal ski time, days on snow or proof of ability. Your account may have booked for other people. The assistant must not infer your skills from those numbers.
This automatic context excludes contact details, date of birth, health information, free-text biographies, payment data, family profiles and other people's records. It supplies no real-account saved preferences to guests, authorised fictional demos or support impersonation sessions. Do not put unnecessary sensitive information in chat.
To correct saved learner preferences, open your account's skiing profile and save the relevant changes. You can also describe what you need for this conversation without changing the saved profile. Other roles use their appropriate account controls.
Saved information is read for the request or a preferences-tool refresh. It is not model training, and a change does not rewrite an answer already given. See the Privacy Policy and conversation guide.
Still need help with this article?
Send the article URL and the step that needs more explanation.