AI GOVERNANCE
How we govern our AI.
CompliGUARD uses AI to help health brands assess their marketing content against TGA · AHPRA · FSANZ · ACCC. This page explains how we govern that AI responsibly — and what it means for you.
HOW WE OPERATE
Six principles. One standard.
CompliGUARD has aligned its AI governance practices with the Australian Government Guidance for AI Adoption (AI6 Framework, October 2025). Here is what that means in practice.
Clear accountability
A named AI governance owner is responsible for every AI system CompliGUARD uses. Our AI systems are documented in an internal register, reviewed annually, and subject to defined controls.
Understood impacts
CompliGUARD's AI use is classified as lower risk — B2B context, advisory outputs only, no automated enforcement decisions. We have formally assessed who could be affected and how.
Managed risks
AI-specific risks are tracked in a formal risk register with defined controls and review dates. We monitor TGA, AHPRA, FSANZ, and ACCC for regulatory changes that could affect assessment accuracy.
Transparent by design
We disclose our AI use in our Terms of Service, Privacy Policy, and Sub-Processors Notice. Our assessments carry an AI Limitations notice on every report. This page is part of that commitment.
Tested and monitored
Every material change to our AI model or assessment methodology triggers a structured test run against a set of known-outcome compliance cases. We audit live assessment quality quarterly.
Human control
No compliance decision should be made solely on the basis of a CompliGUARD assessment. Human review is required — by contract, by design, and by every report we deliver.
DATA PROTECTION
Your content is assessed. Not stored.
CompliGUARD's assessment engine is powered by Claude, developed by Anthropic. When you submit marketing content for review, it is transmitted to Anthropic's API, assessed, and returned as your compliance report.
Anthropic operates under a Zero Data Retention policy for API usage — your content is not stored, logged, or used to train AI models after your assessment is returned. CompliGUARD does not use your submitted content for any purpose other than generating your assessment.
"Zero Data Retention means your marketing content is processed for your assessment and nothing else. It is not retained, analysed further, or used to improve any AI model."
Full details are in our Privacy Policy and Sub-Processors Notice.
HUMAN OVERSIGHT
AI assessments are advisory. Always.
Every CompliGUARD assessment is an advisory output. It is not a legal determination, a regulatory ruling, or a guarantee of compliance. A qualified person must review every assessment before any compliance decision is made.
This requirement is written into our Terms of Service, embedded in every assessment report, and enforced in our product — the dashboard requires acknowledgement of human review before reports are downloaded or shared.
The TGA's March 2024 guidance explicitly endorses AI-assisted compliance review with human oversight as the appropriate model for health marketing compliance. CompliGUARD is designed around that model.
Questions about our AI governance?
We welcome questions from customers, enterprise prospects, and regulatory stakeholders.
Privacy and data questions
privacy@compliguard.com.auAssessment quality or AI feedback
support@compliguard.com.auCompliGUARD is compliance infrastructure, not legal advice. Always engage qualified regulatory counsel for complex TGA matters. We are committed to accuracy, transparency, and continuous improvement as TGA, AHPRA, FSANZ, and ACCC guidance evolves.