AI & Voice Terms

1. AI output is not advice

AI features generate output automatically. Output can be incomplete, out of date or simply wrong, and the same question can produce different answers. Nothing an AI feature produces is legal, tax, accounting, insurance, medical, safety or engineering advice.

You are responsible for reviewing AI output before you act on it or send it to a customer. That includes prices, quantities, taxes, totals, scheduling, addresses, scanned receipt amounts and any message drafted for you. Cervan is not responsible for a decision you make based on AI output.

2. Actions the assistant takes

The assistant can create and change records in your account when you ask it to, for example drafting an estimate, booking an appointment or preparing a message. Those actions are treated as your actions. Confirm anything that sends a message, charges a customer or changes money amounts before you approve it.

3. Voice, microphone and recording

Cervan Voice listens only while you have voice enabled and, where wake-word activation is on, while it is listening for its wake word. Audio is transmitted to our speech providers to convert speech to text and text to speech.

Do not use Cervan Voice to record a phone call, a meeting or any other conversation with another person unless you have the consent every participant's jurisdiction requires. Several states, including California, require all parties to consent. You are solely responsible for obtaining that consent.

Keep sensitive information out of voice and chat input, including full card numbers, bank credentials, government identification numbers and health information.

4. Providers and your data

AI features are delivered through third-party model and speech providers listed in our Subprocessors page. What you submit, plus the account context needed to answer, is sent to those providers to generate a response. We do not sell your data and we do not authorize providers to use your business data to train their public models.

Handling of personal information belonging to your customers is governed by our Data Processing Addendum and Privacy Policy.

5. Fair use and limits

AI usage is metered by plan. Daily request limits and monthly voice minute allowances apply and are described in our Fee Disclosure. When an allowance is used up, the feature pauses until the next cycle or until you add an allowance pack. We may throttle, suspend or change AI features to protect the Service, control abuse or comply with a provider's rules.

Do not use AI features to generate unlawful, deceptive, harassing or infringing content, to impersonate another person or business, to reverse engineer a model, or to build a competing service.

6. No warranty for AI features

AI features are provided "as is" and "as available." We do not warrant that output will be accurate, complete, current, uninterrupted, or fit for a particular purpose. The disclaimers, liability cap and indemnity in our Terms of Service apply in full to AI and voice features.