Artificial Intelligence Policy

1. Purpose and Scope

This AI Policy outlines SQL Insight Solutions' commitment to the safe, ethical, and responsible use of artificial intelligence (AI) in delivering our SQL Server database management services. It applies to all AI systems and tools we deploy, develop, or integrate into our service offerings.

This Policy is aligned with:

• Australia's AI Ethics Principles (2019)
• The Australian Government's Guidance for AI Adoption (October 2025)
• The Voluntary AI Safety Standard's 10 Guardrails
• Australian Privacy Principles
• Relevant sector-specific requirements for healthcare and financial services clients

2. Our AI Philosophy

SQL Insight Solutions integrates AI to enhance—not replace—human expertise. AI assists our team in delivering more comprehensive, accurate, and timely database services. However, critical decisions affecting your systems are always subject to human review, oversight, and approval.

Core Principles:

We commit to:

• Human-Centred AI: AI augments human expertise; humans remain accountable for all decisions
• Transparency: Clear communication about how AI is used in our services
• Security: Rigorous protection of data processed by AI systems
• Continuous Improvement: Regular review and refinement of our AI practices

3. AI Systems We Use

3.1 AI Platforms

We utilise the following AI technologies:

3.2 How AI Supports Our Services

AI is integrated across our service offerings:

Database Health Checks:

• Automated analysis of performance metrics
• Pattern recognition for potential issues
• Risk scoring and prioritisation

Performance Tuning:

• Query analysis and optimisation recommendations
• Index suggestions based on workload patterns
• Resource utilisation analysis

Managed Services:

• Anomaly detection in database metrics
• Predictive alerting for potential issues
• Trend analysis and capacity planning

Migrations and Upgrades:

• Compatibility analysis
• Risk assessment
• Migration planning assistance

Consulting:

• Architecture analysis
• Best practice recommendations
• Documentation generation

4. Human Oversight and Control

4.1 Human-in-the-Loop

All AI-assisted analysis and recommendations undergo human review:

4.2 Intervention Capability

• All AI processes can be paused or terminated by authorised personnel
• Clients can request AI-free service delivery for specific engagements
• Critical changes to production systems require explicit human authorisation

4.3 Decision Authority

AI systems at SQL Insight Solutions:

• CAN: Analyse data, identify patterns, generate recommendations, create draft documentation
• CANNOT: Execute changes on production systems, access data beyond authorised scope, make final decisions on client systems

5. Data Handling and Privacy

5.1 Data Processing Principles

When AI systems process client data:

Minimisation: We process only the minimum data necessary for the specific analytical purpose.

Anonymisation: Where possible, data is anonymised or aggregated before AI processing.

Segregation: Each client's data is logically segregated; one client's data is never used to benefit another client without explicit consent.

No Training: Client data is not used to train third-party AI models. Our use of AI platforms is limited to inference (analysis) only.

5.2 Data Categories and AI Processing

5.3 Third-Party AI Providers

We select AI providers who maintain robust security practices. Our primary AI providers (Microsoft Azure and Anthropic) hold independent security certifications including SOC 2 and ISO 27001. Our agreements with these providers include:

• Data processing terms compliant with Australian privacy law
• Prohibitions on using our data to train their models
• Data deletion upon request
• Appropriate technical and organisational security measures

SQL Insight Solutions' own security practices are informed by industry frameworks including the Australian Cyber Security Centre's Essential Eight and ISO 27001 principles, though we do not currently hold independent certification.

6. Sector-Specific Compliance

6.1 Healthcare Clients

For clients in the healthcare sector, we additionally:

• Comply with AHPRA guidance on AI in healthcare settings
• Ensure AI does not make clinical decisions
• Maintain clear accountability for all AI-assisted work
• Support TGA compliance for any AI that may interact with medical device software
• Provide transparency about AI limitations
• Ensure data sovereignty with Australian-based processing where required

6.2 Financial Services Clients

For APRA-regulated entities and financial services clients:

• Support compliance with CPS 230 (Operational Risk Management)
• Maintain alignment with CPS 234 (Information Security) requirements
• Provide documentation supporting vendor risk assessments
• Enable appropriate oversight of AI-assisted services
• Support third-party audit requirements
• Maintain service continuity practices consistent with prudential standards

7. Risk Management

7.1 AI Risk Assessment

We conduct risk assessments for AI deployment covering:

• Accuracy and reliability of AI outputs
• Potential bias in recommendations
• Data security and privacy implications
• Impact of AI errors on client systems
• Dependency on AI provider availability

7.2 Known Limitations

We acknowledge AI limitations:

• AI can produce incorrect or "hallucinated" outputs
• AI recommendations may not account for all business context
• AI performance depends on data quality
• AI systems may have biases from training data

Mitigation: Human expertise reviews all AI outputs; recommendations are tested before implementation; we maintain non-AI fallback procedures.

7.3 Incident Response

AI-related incidents are managed through:

• Immediate escalation to senior technical staff
• Root cause analysis
• Client notification where material
• Corrective action implementation
• Policy/procedure updates as needed

8. Transparency and Explainability

8.1 Client Communication

We commit to:

• Informing clients when AI is used in their services
• Explaining the role AI plays in analysis and recommendations
• Providing non-AI alternatives upon request
• Answering questions about our AI practices

8.2 Reporting

AI-assisted deliverables will clearly indicate:

• That AI tools were used in the analysis
• What human review was conducted
• Any limitations that may apply

9. Accountability

9.1 Governance

• AI Accountability: The Principal Consultant is accountable for AI governance
• Technical Oversight: Senior Database Administrators review AI-assisted work
• Policy Review: This policy is reviewed annually or when significant changes occur

9.2 Records

We maintain records of:

• AI systems deployed and their purposes
• Significant AI-assisted analyses
• Human reviews and approvals
• Any AI-related incidents or issues

10. Ethical Commitments

Aligned with Australia's AI Ethics Principles, we commit to:

1. Human, Societal and Environmental Wellbeing: AI use should benefit our clients and not cause harm
2. Human-Centred Values: Respecting human rights, diversity, and individual autonomy
3. Fairness: Ensuring AI systems are inclusive and accessible
4. Privacy Protection and Security: Safeguarding data throughout the AI lifecycle
5. Reliability and Safety: Ensuring AI operates reliably as intended
6. Transparency and Explainability: Being clear about when and how AI is used
7. Contestability: Enabling challenge of AI-influenced decisions
8. Accountability: Taking responsibility for AI outcomes

11. Client Rights

Clients have the right to:

• Know: Be informed when AI is used in their services
• Object: Request services without AI involvement
• Access: Understand how AI analysis was conducted
• Challenge: Question AI-assisted recommendations
• Escalate: Raise concerns about AI use to senior leadership

12. Continuous Improvement

We commit to:

• Staying informed of evolving AI regulations and best practices
• Participating in industry discussions on responsible AI
• Updating this Policy as the regulatory landscape evolves
• Training our team on responsible AI practices
• Engaging with the Australian AI Safety Institute (when established)

13. Policy Updates

This Policy will be reviewed:

• Annually as a minimum
• When significant changes to AI regulations occur
• When we introduce new AI systems or capabilities
• Following any significant AI-related incident

Material changes will be communicated to clients.

14. Contact

For questions about our AI practices: