AI use policy
What the AI features do, what data they use, and the human checks around them.
Last updated 18 October 2026
Drafts, never decisions
AI output is always labelled as a draft. Summaries, curriculum drafts and suggested actions are not official until a person with the right role reviews and approves them. AI can miss context or misread records, and people remain responsible for every record.
What data is sent
An AI feature receives only the records the signed-in user is allowed to see and that are needed for the request. Drafts list their sources so a reviewer can check every point.
Providers and training
Requests go to model providers through Vercel AI Gateway. We do not use customer records to train our own models, and we will only use providers whose terms do not allow training on API data.
Limits
The AI assistant does not give legal, medical or safeguarding decisions. If a child may be at risk, follow your safeguarding procedure.
Questions about this document? Email teamsoftwaresolutiongroup@gmail.com.