Skip to main content

Open navigation menu

Beta policy — legal review pending

AI Data Use

AI is optional. Manual direction scoring and every manual path feature work without it. This notice applies only when you explicitly enable and request an AI assist.

Effective July 14, 2026

Beta policy — legal review pending
This document is a good-faith Beta disclosure, not a claim of legal approval. It must be reviewed by qualified U.S. and Canadian counsel before general availability or any paid launch.

Current Beta provider and transfer

The Beta currently plans to use the DeepSeek API for requested AI assists. Relevant text may be processed outside the United States or Canada, including in the People's Republic of China.

DeepSeek's published privacy information describes its own retention, model-improvement, and opt-out practices. This Beta does not promise that provider-side training, retention, or international transfer is disabled. Do not enable AI if those terms are unacceptable or if your content is sensitive.

What may be sent

For each request, the server should send only the task instruction, selected direction, target field, relevant nearby journey text, and source references needed to create the requested suggestion. It should not send payment data, identity-provider tokens, unrelated journeys, or the complete account by default.

  • Goal wording for a requested rewrite
  • Selected current facts for extraction or summarization
  • Relevant inventory, gap, milestone, or next-action fields for a focused suggestion
  • Safety and structured-output instructions needed to validate the result

What AI does not decide

AI does not calculate or rank direction scores, grant access, diagnose personality or mental health, complete a milestone, or directly overwrite formal journey data. Deterministic rules calculate assessment results. You must review and explicitly Apply a suggestion before it becomes a new journey revision.

Sources and uncertainty

Each proposal item is labeled as your fact, a system summary, a suggestion to consider, or a hypothesis to verify. It includes source references and a target field. Generated text may be wrong or invented; verify it before applying.

Storage and logs

Application logs and product analytics must not contain prompts, model responses, or user-authored free text. A successful schema-valid proposal is stored so you can review it. Ignored or expired proposals are retained for up to 30 days. An applied suggestion becomes part of the journey revision and follows journey retention.

Provider systems may retain or use request content under their own terms. Provider retention is separate from the application's 30-day proposal rule.

Promotional Beta assists

The first successful Beta access activation automatically grants 10 promotional AI assists once per account. The balance is shared across journeys and does not expire during the Beta. A successful and persisted group of schema-valid suggestions uses one assist; failures, timeouts, safety blocks, and invalid structures use none. Regenerating a new valid group uses another assist. Review, edit, ignore, and apply actions are free.

AI assists cannot be bought or topped up in the current Beta and have no cash value. Manual features remain available when the balance reaches zero.

Safety boundary

AI output is not professional advice. Requests indicating an immediate self-harm or suicide crisis should be blocked before generation, should not use an AI assist, and should show official U.S. or Canadian crisis resources. In immediate danger, call 911 or go to the nearest emergency department.

Consent and your choice

AI remains off until you accept the current version of the AI notice. Declining AI does not block manual planning. You may stop making AI requests at any time; previously applied text remains ordinary journey content until you edit or delete it. Account export and deletion controls apply as described in the Privacy Notice.