Cavatim AI data-use statement for planner recommendations and agents

AI feature scope

Cavatim uses AI to assist with planning, summaries, reflections, task predictions, goal decomposition, and agent-drafted schedule changes. AI outputs are advisory and the user remains responsible for decisions.

Planner context AI may use

• Goals, deadlines, planner blocks, recurring tasks, workflows, preferences, capacity signals, reflections, journal entries, and notepad context selected by the product flow.
• Connected calendar or mailbox context only when the user has connected that provider and the feature needs that context.
• Support trace and diagnostic context only when the user submits a support request or audited workflow that includes it.

AI data limits

• AI helpers must not intentionally reveal another user's planner items, titles, prompts, or task identifiers.
• Analytics events must not include prompts, generated private planner content, passwords, API keys, reset codes, MFA codes, or excessive planner text.
• Agent planning changes must preserve source linkage and use preview/apply pathways for material planner edits.

Model improvement and provider handling

Cavatim currently sends enabled AI requests to the OpenAI API. OpenAI states that API inputs and outputs are not used to train its models by default unless the API customer explicitly opts in to share data; Cavatim does not intentionally opt production planner data into model training.

OpenAI states that default API abuse-monitoring logs may retain customer content for up to 30 days unless a legally required longer period applies. Cavatim has not claimed Zero Data Retention approval, so users should assume this default provider retention applies when they use AI features.

Cavatim does not use one customer's raw planner content to populate another customer's recommendations. Any future shared or aggregate prediction model must use de-identified or aggregated behavioural patterns and pass privacy review before production use.