1. Scope
These Terms apply to all AI-assisted features ("AI Features"), including task drafts, the assistant, agents, business profile filling, ad copy, e-mail, image and video generation and the CRM assistant.
2. How it works
2.1. AI Features send your inputs (prompts, business profile, records you select) to third-party model providers to generate an answer. Current providers are listed on the Subprocessors page (e.g. OpenAI, Google Gemini, Anthropic).
2.2. To reduce cost and unnecessary calls, many operations are first done with rules; AI is called only when needed. Answers for the same request may be cached for at most 14 days.
2.3. If you connect your own API key (BYOK), requests are made through your account; that provider's terms and fees apply to you.
2.4. To prevent abuse, comply with the law and protect the Service, we may filter, refuse or log prompts and outputs. Security logs follow the retention period in the Privacy Policy. The cache is at most 14 days.
3. Your data
3.1. Your inputs and Output are not used to train general AI models. We work with providers under business terms that do not use API data for model training.
3.2. Providers may retain data for a limited period stated in their own policies to monitor abuse.
3.3. We recommend not entering special categories of personal data, trade secrets or information subject to confidentiality obligations into AI Features. The cache, and a provider's own retention period, are not a separate confidentiality commitment beyond the Privacy Policy and the Data Processing Agreement. Do not enter as a prompt information you cannot share with a subprocessor.
4. Limits of Output
4.1. AI may produce incorrect, incomplete, outdated, misleading or biased output and may "invent" facts, sources or figures that do not exist.
4.2. Output is only a draft. Reviewing it before publishing, sending, using it in an ad or relying on it for a decision is your responsibility.
4.3. Output is not legal, financial, tax, medical or investment advice.
4.4. The same input may produce different output at different times; similar output may be produced for other users, and this does not create exclusivity.
5. Rights and responsibility
5.1. Rights in Output belong to you to the extent permitted by applicable law and provider terms. You acknowledge that AI-generated content may not be protected by copyright in some jurisdictions.
5.2. It is your responsibility to check that Output does not infringe third parties' copyright, trademark, personality or other rights, and that it complies with advertising, consumer and commercial messaging law. We do not promise that Output is original, accurate or non-infringing.
5.3. You may generate content containing a real person's image, voice or name only with that person's express permission.
5.4. Where platforms or law require it, labelling AI-generated content as such is your obligation (e.g. Meta's and Google's policies on AI-generated content).
6. Usage limits
Monthly usage limits may apply to AI Features depending on your plan. In case of abuse or unusual load, we may slow down or stop requests.
7. Prohibited use
In addition to the Acceptable Use Policy, AI Features may not be used for purposes contrary to the providers' usage policies, with inputs designed to bypass safety measures (jailbreaks), or to collect output in bulk to train a competing model.
8. Liability
We are liable for damage arising from use of Output within the limits in the Terms of Service, and to the extent those limits are enforceable. Consequences of publishing, sending or presenting Output to a third party as advice without reviewing it are yours. Liability for intent, gross negligence and liability that statute does not allow to be limited remains reserved.
Version: 2.0.