Legal
AI use policy
How Kim uses artificial intelligence, in plain language.
Draft pending validation. Text written by induSmart.digital; it must be reviewed by a lawyer and by INNOVANDIA, SLU, which will validate it in writing before launch. The Companies Register entry number is still missing.
Last updated: 12 November 2026
Kim suggests, it doesn't act
Kim prepares replies, tasks and appointments for you. By default, it only suggests: you approve, edit or discard, and what you approve can be undone. Nothing is sent or deleted without your confirmation.
Only with your consent
Kim only uses AI if you've given your consent, and you can withdraw it whenever you want from the settings. On the Free plan, Kim doesn't use AI: it sorts using automatic rules and doesn't send your text to any AI provider.
Which models
Kim uses AI models from Anthropic. On paid plans, with AI; on the higher plans, also with advanced AI, a limited number of times a month. You'll always see which level you're using and how much you've used.
What we send and what we don't
Only the text needed to prepare the suggestion you ask for. We don't use your data to train general-purpose models. Kim can improve with anonymous, aggregated signals, never with your content.
Your key, if you like
You'll be able to connect your own key from an AI provider (BYOK). In that case, the processing is done by that provider under its own terms. The key is stored encrypted and Kim never sees it.
Your AI assistant, with permission
If you connect Duokim to Claude or ChatGPT (MCP server), your assistant can only read what you can already see; write, only if you switch it on; delete or send, never.
Limitations
AI makes mistakes. Always review Kim's suggestions before approving them.