CoachVana uses AI in two ways: features we build into CoachVana, which run on Anthropic's models, and AI assistants such as Claude or ChatGPT that a coach can connect to their own account. This page says what each one receives, what it never receives, and how to turn it off. It is part of our Privacy Policy.
Never sent to any AI, whoever asks
- Data from a watch, ring, smart scale or health app: Apple Health, Health Connect, Oura, Fitbit, WHOOP and Garmin.
- Events from a coach's Google Calendar.
- Passwords and payment card details.
No AI model is trained on your data by us, and our AI provider does not train on what we send it.
AI features in CoachVana
These features use AI. Each one runs only when a person uses it:
- Estimating a meal from a photo or a description, and reading a nutrition label.
- Turning a document, spreadsheet, photo or pasted text into a program, meal plan, recipe or form, for a coach or for a client logging on their own.
- Building a program or a meal plan for a coach.
- Suggesting changes to a client's program from the training they logged.
- Writing a coach's outreach messages.
- Summarizing a client's recent week for their coach, drafting a reply in a coach's client thread, and Coach Copilot, which drafts a check-in message for a coach.
What is sent.Only what that feature needs. For a summary, a reply draft or a Copilot draft, that can include a client's name (a Copilot draft uses the first name only), logging totals such as workouts and calories, recent check-in answers, and recent messages between the coach and the client. For a program suggestion, the sets, reps and loads the client logged. Where a draft quotes messages, the model is told they are words people wrote, to read as data and never to follow as instructions.
What is never sent, besides the list above: body measurements, photos and videos, a coach's private notes, health and intake forms, email addresses and payment details.
Who processes it. Anthropic runs these features as our service provider. Under its commercial terms it may not train its models on what we send, and it deletes inputs and outputs within 30 days, except where it must keep them longer for the law or to enforce its usage policy.
A person decides. Nothing an AI writes is sent to anyone, assigned or applied until a person chooses to. AI output can be wrong: it is an estimate or a draft, not medical, dietetic or professional advice, and it is not reviewed by a professional before you see it. We never use AI to make a decision about you that has a legal or similarly significant effect.
AI features are optional and metered. A coach who doesn't use them sends nothing.
AI assistants a coach connects
A coach can connect an AI assistant, such as Claude (made by Anthropic) or ChatGPT (made by OpenAI), to their CoachVana account, so they can ask it about their coaching business and have it prepare work. The coach connects it from the assistant and approves it on a CoachVana page that shows which assistant is asking. Only coach accounts can connect one.
What it can read:what the coach sees in CoachVana, such as their clients' names and status, the programs, plans, notes and messages the coach wrote, their calendar, their leads and their business figures.
What a client logged is theirs to share. An assistant reads what a client logged or wrote (workouts, typed weigh-ins, food, check-in answers, messages they sent) only if that client has said yes in their own CoachVana app. Each client is asked once, and can change their answer at any time in Settings. Until a client says yes, an assistant sees their name and status and what the coach wrote, and nothing else of theirs. Email addresses and phone numbers are left out unless one is needed for what the coach asked.
Enforced by our database.The limits above, and the list of what is never sent, are applied by CoachVana's database to every request an assistant makes, not left to the assistant.
What it can do: prepare messages, replies and changes for clients, which happen only after the coach approves each one, and save things only the coach sees, such as notes, tags and drafts. It cannot delete anything, move money or change prices.
Where the data goes.What the assistant reads to answer the coach is sent to its provider, Anthropic for Claude or OpenAI for ChatGPT, which processes it under its own terms and privacy policy and the coach's own settings with it, including whether conversations may be used to improve its models. That is between the coach and that provider. We send data only to answer the coach's request.
What we keepis a record of each request the assistant makes (which tool, when, whether it worked and how long it took), never the conversation and never what was asked, for 90 days, and each action waiting for the coach's approval, which expires after 15 minutes and is deleted a day later. Both go with the account.
Your controls
- Clients:change whether an assistant may read what you log in Settings, with Share my logs with AI assistants. Turning it off takes effect on the assistant's next request.
- Coaches:disconnect an assistant at any time in Settings, under Connected AI assistants. It stops at once. Remove CoachVana in the assistant's own settings too.
- Anyone: AI features run only when you use them. Deleting your account deletes what we hold, as our data deletion page describes.
Questions about AI and your data: support@coachvana.com.