AI Use and Safety Policy

1. Where we use AI

We may offer optional AI-assisted features, including Start Here routing, Quick Help summaries, grounded website answers, scam-message explanations, business cyber self-check explanations and customer-safe My Tech Help status explanations, explanations of human-approved quotes and scope, and plain-English explanations of deterministic scam-recovery plans. We may add, change, suspend or remove an AI feature where it is unreliable, unavailable, unsafe or no longer appropriate.

2. Optional use and non-AI alternatives

AI features are optional unless a page clearly says otherwise. Existing pathway cards, deterministic checkers, forms, telephone support and human review remain available. Refusing an optional AI feature does not remove your ability to request ordinary help.

3. AI outputs are not final decisions

AI output may be incomplete, inaccurate, outdated, misleading or unsuitable for your circumstances. It is general guidance only. You must not treat it as a diagnosis, binding quote, booking approval, warranty decision, provider approval, referral-reward approval, legal advice, financial advice, cyber-forensics report, compliance assessment, certification, insurance approval, government endorsement or guarantee of security.

4. Human review and important decisions

A suitably authorised person must review important decisions involving price, paid work, bookings, warranties, refunds, disputes, provider approval, referral rewards, legal rights, financial loss, account compromise, safety, privacy or security. We may decline to provide an AI response and direct you to a person or official service.

5. Information sent to AI service providers

When you actively use an AI feature, the text or structured selections needed to generate the response may be sent through our server to third-party AI infrastructure, currently including Google Gemini. These providers may process information outside Australia and their own terms and privacy practices apply to their processing.

Google Gemini processing is governed by Google’s applicable API terms and privacy practices. We do not promise that every provider has identical retention, logging, location or training practices. Provider and model arrangements may change.

6. What you must not submit

You should remove names, account numbers, addresses and other identifying details before using an AI feature. Automated redaction reduces some risk but cannot identify every secret or personal detail.

7. Consent and control

Where a feature asks for consent, the AI request will not be made until you take the required affirmative action. Consent applies to that request and does not create permission for unrelated marketing, publication or automatic decision-making. You may stop before submission and use the non-AI pathway.

8. Storage, logs and security

We design customer AI endpoints to use server-side calls, input limits, rate limits, CSRF and same-origin protections, redaction and safe output rendering. Where a feature states that pasted text is not stored in our application database, we do not intentionally save that text as a customer record. Technical infrastructure and external providers may still create security, abuse-prevention, operational or legal logs under their own policies. Internet transmission and automated redaction cannot be guaranteed completely secure.

9. My Tech Help limitations

The My Tech Help AI assistant may explain only customer-safe information already linked to your authenticated account. It does not access technician-only or admin-only notes, does not change a request, and cannot approve quotes, bookings, payments, warranties, refunds, disputes, provider decisions or completion dates. The visible customer record and authorised staff remain the source of truth.

10. Scam and cyber limitations

An AI scam explanation cannot prove that a message is safe or fraudulent. Verify independently using an official app, a website address you type yourself, or a known phone number. Do not click, reply, pay, install remote-access software or share codes because an AI response appears reassuring.

Business cyber scores remain deterministic where stated. AI only explains the recorded result. The feature does not certify Essential Eight maturity, formal compliance, insurance eligibility or guaranteed security.

11. Prohibited use

You must not use our AI features to facilitate crime, fraud, impersonation, harassment, surveillance, unauthorised access, malware, credential theft, deceptive content, discrimination, exploitation, self-harm, sexual abuse material, or infringement of privacy, intellectual property or other rights. You must not probe, bypass or interfere with safety controls, rate limits, prompts, access controls or service infrastructure.

12. Children and vulnerable users

Children and vulnerable persons should use AI-assisted safety or support features with a trusted adult, guardian, support person or authorised worker where appropriate. AI must not replace safeguarding, emergency assistance or qualified professional support.

13. Availability and changes

AI services may be delayed, unavailable, rate-limited or changed without notice because they depend on third-party infrastructure and model availability. We may provide a non-AI fallback instead. We may update this policy as features, providers, laws and safety practices change.

14. Consumer rights and liability

Nothing in this policy excludes, restricts or modifies any right or remedy that cannot lawfully be excluded, including rights under the Australian Consumer Law. Subject to those rights, you remain responsible for checking AI output before relying on it and for obtaining qualified advice for important, urgent or high-risk matters.

15. Questions, corrections and complaints

Contact us through the support page or the contact details published on this website if an AI answer appears unsafe, incorrect, discriminatory, privacy-invasive or misleading, or if you want a human review. Privacy requests are handled under our Privacy policy.

Related policies