How the AI Coach works.

The Coach is a career-growth co-pilot. It helps you turn honest self-assessment into forward motion — reframing setbacks, preparing for interviews, working through a promotion case, planning your next 90 days, or thinking about a career pivot.

It uses the frameworks we’ve published across the site, your latest assessment results, and the direction you’re steering the conversation toward. It’s a mentor’s read, not a decision-maker.

Every time you send a message, the Coach receives:

• The message you just typed and any file you attached to it (text is extracted server-side).
• Recent messages from the same conversation, so it can follow the thread across turns.
• Your latest assessment scores (per category, plus overall) — so it can meet you where you are.
• Your resume text, if you’ve uploaded one, so it can tailor advice to your actual experience.
• The coaching mode you’ve selected (career, leadership, interview prep, resume review).

• Your password, or anything derived from it. Passwords are bcrypt-hashed and never leave our server.
• Your payment or Stripe details. Billing is handled by Stripe directly.
• Your IP address, browser, device fingerprint, or location.
• Any other member’s data. Every prompt is scoped to your account.
• Admin-only fields — roles, audit logs, internal notes.

The Coach runs on Anthropic's Claude models — currently claude-sonnet-4-5-20250929, accessed through an enterprise LLM proxy operated by our platform partner Emergent.

Training. Anthropic does not use inputs or outputs from its commercial API products to train its models by default. We do not fine-tune any model on your data, and we do not send member data to OpenAI or Google for coaching. If that ever changes, we will say so on this page — clearly, and before it takes effect.

Retention at the provider. Under Anthropic’s standard API terms, inputs and outputs are generally deleted from Anthropic’s systems within 30 days, subject to applicable terms and limited exceptions (for example, abuse investigations or legal holds). Unbreakable One does not currently have a separate zero-retention agreement with Anthropic.

Separate from Anthropic’s temporary processing, Unbreakable One stores your coaching conversations on our own servers so you can pick up where you left off, so we can show you a real history rather than a moment-to-moment forget-everything experience, and so the Coach can carry context across turns.

• Where it lives: our MongoDB database, in the `chat_messages` collection, scoped to your `user_id`.
• How long we keep it: for the life of your account, unless you delete it sooner. We do not yet run automatic time-based purges — a proposed 24-month rolling retention is under review.
• Who can access it administratively: only a small number of Unbreakable One administrators, and only when necessary for support, safety, or legal compliance. Admin views are logged in an internal audit trail.
• How it personalizes future coaching: recent messages from a session are re-supplied to the Coach so it can reason across turns; deleted conversations stop influencing personalization on next send.

• Delete an individual conversation from the Privacy Dashboard (/app/privacy) — instant, no waiting period.
• Clear every conversation on record from inside the Coach.
• Deleting your account (from the Privacy Dashboard) purges conversations along with everything else on our side.

Deletion on our side removes the conversation from your account immediately. Any copies still held by Anthropic within their 30-day retention window fall out on their normal schedule.

The AI Coach is not a licensed therapist, financial advisor, doctor, or lawyer. It shouldn’t be used for clinical decisions, medical intake, legal filings, or investment advice. For those, please talk to a qualified human.