Straight answers,
including the noes.

No. Certification follows an independent audit by an accredited body. SafetrexIQ builds the management system that audit examines.

No. Generated frameworks, policies and assessments are drafts and templates for a competent person to review and approve.

No. It gives a rules-based reportability verdict with the regulation cited. Deadlines and filing remain your duty.

United Kingdom first. Statutory logic lives in a jurisdiction pack, and documents cannot be approved until a named competent person has signed off the pack.

Your records are stored in France. If your organisation's AI mode allows it, AI drafting requests go to OpenAI, which may process them outside the UK and EU, including in the United States, after recognised names and contact details are replaced with placeholders. Health information we detect keeps the whole request on our own model in France. Detection uses word lists and rules, so it can miss wording they do not cover. It is a safeguard, not a guarantee. Occupational-health records are kept in separate tables behind an OH-role gate.

Yes. Off is one of four AI modes, set per organisation by an owner, admin or health and safety manager. In Off mode AI drafting and the assistant are switched off, so you draft by hand. Search still indexes records with our own on-server model; nothing is sent to OpenAI. Everything else keeps working.

Off: AI drafting and the assistant are switched off; search still indexes records on our own model. Sovereign: nothing goes to OpenAI and all AI runs on our own model in France. On current hardware that is slow: Ask Trex still answers, more slowly, but document drafting (management system, risk assessments, investigation write-ups, site recommendations) does not complete and must be written by hand. Restricted: only anonymous totals and trends may go to OpenAI; no current feature sends totals, so today it behaves like Sovereign. Standard, the default for new organisations: requests may be processed by OpenAI (OpenAI OpCo, LLC), including in the United States; recognised names and contact details are replaced first, and detected health information stays on our own model.

Only what your organisation's AI mode allows. In Standard mode, drafting requests may go to OpenAI after names recorded in your workspace, names with a title, email addresses, phone numbers, postcodes and National Insurance numbers are replaced with placeholders. Other names typed into free text may not be recognised. Health information is detected automatically and kept on our own model. Detection uses word lists and rules, so it can miss wording they do not cover. If your organisation cannot accept that risk, choose Sovereign or Off. Search indexing runs on our own servers and never goes to OpenAI.

No. Each organisation's records carry its id, and the server returns only the signed-in organisation's records.

Anyone with the QR code or link, with no login and no app. They get a reference code and a page showing any update your team posts.