AI Use & Transparency Statement

1. Scope of this statement

This statement applies to:

• AI tools and services we use internally (for example, to support productivity, research, software development, marketing, and operations).
• AI tools and services we use in client work, including:
  • AI-assisted analysis and reporting
  • AI-assisted marketing, content, and UX work
  • AI-enabled automations and integrations
  • Other AI-supported solutions we design or implement for clients

It does not replace any written agreement we have with you. If there is ever any inconsistency, the terms of our contract, proposal, or statement of work prevail.

2. Our responsible AI principles

Quantum is guided by widely recognised responsible AI and digital ethics principles, including Australian Government guidance on safe and responsible AI. In practice, this means we aim to ensure that our use of AI is:

• Human-centred
  • AI is used to support people, not to replace human accountability.
  • Important decisions are always subject to human oversight and review.
• Fair and non-discriminatory
  • We avoid using AI in ways that intentionally discriminate against individuals or groups.
  • For higher-risk use cases, we consider whether additional checks are appropriate to identify potential bias.
• Privacy-respecting and secure
  • We minimise the personal and sensitive data that is shared with AI tools.
  • We prefer configurations and providers that support strong privacy and security controls.
• Transparent and explainable
  • We aim to be open with clients about where AI plays a material role in our services.
  • We explain, in plain language, how AI features or automations will be used in your project.
• Accountable
  • Quantum remains responsible for the services we provide, even where AI tools are involved.
  • We treat AI outputs as inputs into our professional work, not as unquestioned answers.

3. How Quantum uses AI in our own business

Internally, Quantum uses a range of AI systems and tools to support our day-to-day work. Typical examples include:

• Drafting and editing
  • Generating and improving drafts of emails, documents, reports, presentations, and marketing material.
• Research and summarisation
  • Summarising long documents, extracting key points, and assisting with background research.
• Software development assistance
  • Assisting with code suggestions, refactoring, documentation, and test generation.
• Analysis and insight
  • Supporting analysis of data sets, performance trends, or content for internal decision-making.

In all cases:

• AI outputs are reviewed and, where necessary, edited by Quantum staff before being relied upon or shared.
• We do not rely solely on AI outputs to make important business decisions without human review and judgement.

4. How we use AI in client work

AI is an important tool in helping us deliver efficient and effective work for clients. Depending on the engagement, we may use AI to support:

• Content and creative work
  • Ideation, outlining, and drafting of content (e.g. website copy, ad variants, email flows), which is then refined and tailored by our team.
• Analysis and optimisation
  • Analysing large sets of performance or behavioural data to identify patterns, opportunities, or risk areas for further human review.
• Automation and workflows
  • Designing and implementing AI-enhanced workflows, automations, or assistants that help streamline internal processes and tools for clients.
• Technical implementation
  • Supporting code generation, documentation, or configuration for AI-related features and integrations.

Our guiding position is that AI is used as an accelerator and assistant. Final recommendations, deliverables, and implementations are approved by humans at Quantum, and we remain accountable for the services we provide. We do not allow AI systems, on their own, to:

• Enter into contracts on behalf of clients;
• Make legally or financially binding decisions without human review; or
• Replace any regulatory or professional judgement where human expertise is required.

Where AI plays a material role in a service or deliverable, we will make this clear in our proposals, statements of work, or supporting documentation.

5. Use of client data with AI tools

We treat client data used with AI tools in the same way as any other confidential information.

5.1 Data minimisation

Where possible, we:

• Avoid sending personal or sensitive data to AI tools unless it is strictly necessary for the task and agreed as part of the engagement.
• Use de-identification, pseudonymisation or aggregation where this is practical and still allows us to perform the work.

5.2 Training and model improvement

Where we have control over settings:

• We prefer to use enterprise or API-based AI services configured so that client data is not used to train or improve public models.
• We do not intentionally permit client confidential information to be used to train models that are then made available to other customers.

For any solution where training on client data is part of the design (for example, a private knowledge base or internal chatbot for the client), we will:

• Make this clear in project documentation; and
• Design the solution so that training occurs within an agreed and controlled environment (for example, a client-scoped model or data store).

5.3 Security and confidentiality

Any use of client data with AI tools is subject to the same security and confidentiality commitments that apply under:

• Our Privacy Policy;
• Our Terms & Conditions; and
• Any specific agreement or non-disclosure obligations we have with you.

We select AI tools and configurations with security in mind, but no system can be guaranteed as completely risk-free. We will act reasonably and promptly if we become aware of a security incident involving AI tools that affects your information.

Data location. For engagements that require it, we can process client data on AI infrastructure located in Australia. In particular, we use Amazon Bedrock in the Amazon Web Services Asia Pacific (Sydney) Region, so that client data is processed onshore and is not used by model providers to train their models.

If an AI provider we rely on experiences a security incident affecting client data, we treat it as a data breach under our Privacy Policy and any applicable Data Processing Agreement, and notify affected clients accordingly.

6. Model and tool selection, including higher-risk tools

The AI ecosystem changes rapidly. Quantum therefore uses a variety of AI models and services, and the specific tools we use may change over time as technology, security, and commercial factors evolve.

We generally do not list specific models or vendors in this statement for that reason. Instead, we commit to the following:

• We select AI tools based on a combination of capability, reliability, security, privacy posture, and contractual terms.
• We prefer configurations that minimise data exposure and allow us to control how information is stored and processed.
• We regularly review our AI toolset and may add, update, or remove tools as circumstances change.

Where we form the view that a particular AI tool or provider presents a heightened security or privacy risk, we may:

• Restrict its use to specific low-risk contexts; or
• Prohibit its use entirely for client or internal work.

This includes avoiding tools where there are credible concerns about how data is handled, stored, or shared. From time to time, this may mean we choose not to use certain popular or emerging AI tools or models (including, for example, tools similar to DeepSeek or any service where we consider the security or privacy risk profile to be unacceptable for client work).

We do this to protect both Quantum and our clients, even where a tool might appear attractive from a cost or capability perspective.

7. Intellectual property and originality

AI systems are trained on large data sets and can generate content that is similar in style to material they were exposed to during training. Because of this:

• We use AI as a supporting tool for ideation, drafting, and refinement, not as a sole source of creative or technical work.
• Quantum team members add original human contribution, curation, and editing to AI-assisted outputs before delivering work to clients.

While we act in good faith to avoid infringing third-party intellectual property, neither Quantum nor any AI provider can guarantee that AI-assisted content is entirely unique or free from similarity to existing works.

If you notify us of a credible intellectual property concern relating to AI-assisted work we have provided, we will:

• Review the concern in good faith; and
• Where appropriate, modify or replace the relevant material to address the issue.

8. Accuracy, bias, and limitations of AI

AI tools can be extremely powerful, but they also have well-known limitations. For example, they may:

• Produce information that is incorrect, incomplete, outdated, or misleading;
• Reflect or amplify biases present in their training data;
• Misinterpret ambiguous prompts or contexts;
• Fail to account for the most recent regulatory or market changes.

To manage these limitations, Quantum:

• Treats AI outputs as drafts, suggestions, or inputs, not as definitive answers.
• Applies human review and professional judgement before relying on AI-assisted outputs in client work.
• Uses additional validation for high-impact or higher-risk tasks (for example, cross-checking critical figures, legal references, or regulatory interpretations against authoritative sources).

Even with these safeguards, you should understand that AI-enabled services inherently carry some residual risk. Our contractual terms and disclaimers describe how liability is managed in these situations.

9. Client choices and constraints

We recognise that different organisations have different risk appetites and internal policies for AI. Subject to the scope and commercial terms of the engagement, we will:

• Work with your legal, risk, or compliance teams where AI use is subject to specific internal policies or sector-specific regulation.
• Discuss and document any additional controls you require for AI-related components of a project.
• Where feasible, accommodate reasonable requests to:
  • Avoid particular AI tools or vendors; or
  • Restrict the use of AI to certain parts of the engagement.

In some cases, limiting AI use may:

• Increase the time and cost required to deliver a project; or
• Reduce the scope of what can be automated or optimised.

If that is likely to be the case, we will explain the trade-offs and discuss options with you before work proceeds.

10. Governance, staff training, and continuous improvement

Quantum is committed to improving how we use AI over time. To support this, we:

• Provide staff and contractors with guidance and training on responsible and secure AI use.
• Maintain internal practices for reviewing and approving new AI use cases before they are put into production.
• Periodically review the AI tools we use, as well as our internal guidelines, in light of:
  • Changes in technology;
  • Changes in applicable laws or regulations; and
  • Lessons learned from our own work and from the broader AI ecosystem.

This AI Use & Transparency Statement itself will also be reviewed at least annually, or more frequently if there are material changes in how we use AI or in the legal and regulatory environment.

11. Relationship to other Quantum policies

This statement forms part of the wider set of policies that govern your relationship with Quantum. It should be read together with:

• Our Privacy Policy (how we collect, use, and protect personal information);
• Our Terms & Conditions (contractual terms that apply to our services);
• Our Disclaimer (limitations and risk allocation relating to information on our website and our services).

If you engage Quantum under a specific contract, proposal, or statement of work, that document may contain additional terms governing our use of AI for that engagement.

For clients whose engagement involves us handling personal information on their behalf, our Data Processing Agreement sets out the detailed data-protection terms and prevails over this statement on those matters.

12. Contacting us about AI

If you have any questions or concerns about how Quantum uses AI—whether in our own business or in work we carry out for you—you can contact us using the details on our Contact page or any dedicated email address we specify for privacy or governance matters.

Please include enough detail for us to understand your question or concern (for example, the relevant project, service, or page), and we will endeavour to respond within a reasonable timeframe.