Skip to main content

Legal · BETIHUT-LEGAL-AIT

AI Transparency Notice

Entity
AUTOMAIZE SOLUTIONS LTD
Reg. No.
HE 480609
Effective
18 August 2026
Status
Version 2.0 · Effective

This AI Transparency Notice explains how AUTOMAIZE SOLUTIONS LTD, as the sole current Betihuti operator, provides and deploys the automated features of the Betihuti workplace-safety service, the safeguards applied to them, and the rights available to you. It supplements, and should be read together with, our Privacy Notice.

It is provided in furtherance of the transparency obligations under Articles 13 to 15 and 22 of the GDPR, and of the transparency and human-oversight expectations that apply where automated processing is used in a health-and-safety context.

1.Who is responsible

  1. 1.1Automaize designs, maintains and operates the Betihuti automated features and their audit controls, determines the permitted purposes, supervises the use of outputs and provides the human-review route.
  2. 1.2The Customer that enrolled you defines its own safety requirements, decides which automated features it enables, and remains responsible for every safety decision it takes — including whether you may enter a site, operate an asset or perform a task. The applicable contract and law may assign additional provider or deployer duties; these labels do not reduce any party's statutory obligations.

2.When you see this notice

  1. 2.1We link to this Notice at signup, at the consent step for the safety assistant, from any surface that shows an automated readiness or authorisation outcome, and from the human-review route.
  2. 2.2You do not need to accept an automated feature without being able to read this Notice first.

3.What the automated features do

  1. 3.1Readiness and authorisation evaluation. The Service compares the training, qualifications, medical-fitness status, equipment issue and permits recorded about you against the requirements your employer has set for a site, asset or task, and returns whether a requirement is met, due or expired. This is a deterministic comparison of dated records against stated rules, not a prediction about you.
  2. 3.2The safety assistant. Where your employer enables it, an assistant answers safety questions using that employer's own procedures, method statements, risk assessments and training material. Its answers cite the source document and passage they came from. It can look up your own readiness and explain which requirement is unmet.
  3. 3.3Assisted drafting and summarisation. The Service may transcribe a recorded briefing or spoken report and may propose a summary of an incident report or a draft corrective action. A proposal is never applied on its own: a person with the relevant authority must review and confirm it before it changes any record.
  4. 3.4Where the assistant is asked to do something outside the permitted set, or the underlying feature is not enabled for your employer, it refuses rather than improvising.

4.What the automated features do not do

  1. 4.1They do not score, rank or rate you as a person, do not profile your character or aptitude, and do not produce a suitability or employability assessment.
  2. 4.2They do not diagnose you, interpret clinical records or form a medical opinion. A medical-fitness outcome shown in the Service is the conclusion of a qualified examiner recorded in it, not a conclusion the Service reached.
  3. 4.3They do not make your employer's safety decisions and do not replace supervision, instruction, training or a competent person's judgement.
  4. 4.4They do not use protected attributes, and do not use payment data or unrelated Customer data, as inputs to any evaluation.

5.Legal basis

  1. 5.1We carry out this processing to perform our agreement to provide the Service and, where the processing carries out an occupational health-and-safety obligation, on the bases described in the Privacy Notice, including Article 9(2)(b) GDPR for occupational health data.
  2. 5.2Optional features that are not necessary to the safety programme — including the safety assistant and any recording of a briefing or spoken report — run only after your recorded consent, which you may withdraw at any time without affecting the lawfulness of processing carried out before withdrawal.

6.How we keep the evaluation fair and inspectable

  1. 6.1Everyone assigned the same role at the same site is evaluated against the same published requirement set, and that requirement set is visible to you.
  2. 6.2An outcome names the specific requirement that was not met, together with the record and expiry date behind it, rather than a bare score. You can see why, not just what.
  3. 6.3An outcome changes as soon as the missing evidence is recorded; nothing about a past shortfall persists as a rating.
  4. 6.4Every assistant answer keeps a citation to the document version it relied on, and every automated evaluation keeps an audit trail of its inputs, the rule applied, the model and prompt versions where a model was involved, and the result — so that a review can be grounded in the same evidence.
  5. 6.5Where a requirement is unmet the Service withholds authorisation rather than granting it. A safety feature that fails must fail closed.

7.Human oversight and your rights

  1. 7.1You have the right not to be subject to a decision based solely on automated processing that produces legal or similarly significant effects concerning you (Article 22 GDPR). Being prevented from working can be such an effect, and your employer must provide a person who can review it.
  2. 7.2You may request human intervention, express your point of view and contest any automated outcome. Where a human review changes the outcome, the revised result replaces the automated one in your record.
  3. 7.3If you believe a record behind an outcome is wrong or out of date — a certificate that was renewed, a fitness result that was superseded — you may ask for it to be corrected, and the outcome is recalculated from the corrected record.
  4. 7.4You may access, export or delete your data, withdraw consent where applicable, and exercise your other rights as described in our Privacy Notice.

8.Data used and retention

  1. 8.1An evaluation may use your training and qualification records, medical-fitness status and restrictions, equipment issue, permits, site assignment, presence records, the requirement set your employer has defined, consent records and audit logs.
  2. 8.2The safety assistant additionally uses your question, the conversation you are in, and the passages retrieved from your employer's own safety library. It does not read another employer's data, and it does not read parts of your record that your role and the enabled features do not permit.
  3. 8.3Assistant conversations, transcripts and proposed drafts are retained on a defined retention schedule so that an outcome can be checked if you request a review, after which they are deleted or anonymised. You may request erasure sooner, subject to any retention required by law and to the records your employer must keep by statute.

9.Assistant answers

  1. 9.1The assistant is expected to answer from the cited source rather than to invent a reason, and to say that it does not know when the safety library does not contain the answer.
  2. 9.2Assistant answers are informational. They do not replace your employer's procedures, a permit, a competent person's instruction, an emergency procedure or any legal notice.
  3. 9.3Never rely on an assistant answer in an emergency. Follow your site's emergency procedure.

10.Limitations

  1. 10.1An automated evaluation is only as good as the records behind it. It cannot know about training you received but that was never recorded, or a condition you never declared.
  2. 10.2An outcome of "authorised" means the recorded requirements were met. It is not a statement that a task is safe to perform on the day, and it never displaces a dynamic risk assessment, a pre-use check or your own right to stop unsafe work.

11.Contact

  1. 11.1If you have questions about these features, a review or your data rights, contact privacy@betihuti.com.