How Tickbox uses AI

Last updated: [EFFECTIVE DATE]

We want to be upfront about how AI is used in Tickbox, what happens to your data, and where the limits are.

What the AI does

Tickbox uses a large language model (Anthropic's Claude) for two things: reading an uploaded questionnaire to work out the list of questions, and drafting an answer to each question.

What we send to the AI

To generate drafts, the content of your uploaded questionnaire, the questions we extract, and your relevant saved answers are sent to Anthropic. We send only what is needed to produce the drafts.

Your data is not used to train models

We use Anthropic through its API. Content sent this way is not used to train Anthropic's models.

Drafts, not decisions

Every answer the AI produces is a draft for you to review, edit, and approve. Tickbox does not decide whether you are compliant and does not issue any certification. You are responsible for checking each answer before you send it.

Accuracy

AI can make mistakes or miss context. Treat every draft as a starting point, not a final answer, and correct anything that is not right for your business.

Where processing happens

Some AI processing takes place outside the UK. Where it does, appropriate safeguards apply. See our Privacy Policy for detail on data handling and your rights.

Questions

Anything about our use of AI: [CONTACT EMAIL].