Independent cybersecurity assurance
AI Assurance
Know whether the safeguards around your use of AI work in practice. Independent assessment of cybersecurity controls, information exposure and the evidence supporting responsible use.
Your AI use. Your information. Your accountability.
AI tools can introduce new ways for information to be accessed, shared and acted on. Approving a tool or documenting a policy does not, by itself, show that its safeguards operate as intended.
Sypha assesses the controls protecting your information, systems and use of AI. We examine evidence of access, information handling and agent permissions, identify weaknesses and gaps, build the AI policies and governance frameworks you need, and set out a prioritised maturity roadmap.
We support government bodies, their suppliers, Defence supply-chain businesses and other organisations using AI with sensitive information. The starting point is your actual use of AI and the obligations that apply to it.
For executives
You are accountable for how AI is used with sensitive information, and for what you tell your board, regulators and buyers. The assessment shows whether safeguards approved on paper operate in practice, what remains uncertain, and which decisions follow, such as which use cases to approve, restrict or pause. Your organisation decides how to use AI and implements the safeguards. We agree the scope and a fixed fee once we understand your requirements. Start with a free scoping conversation.
How we can help
Find the right starting point
Not sure where to start?
You do not need to choose an assessment before speaking with us. Start with how your organisation uses—or plans to use—AI, the information involved and the decisions you need to make. In a free 30-minute consultation, we’ll discuss your priorities and clarify a useful next step.
01 / AI adoption readiness assessment
Confirm the requirements, tools and use cases in scope. Assess the controls around adoption, verify a sample in practice and receive a gap register, independent findings and a prioritised roadmap.
02 / AI use-case assurance
For a specific use case, independently review the impact or privacy impact assessment and security risk assessment. Examine whether the proposed safeguards address the identified risks and whether supporting evidence is sufficient.
03 / Ongoing AI assurance
Re-verify safeguards as tools, access and use cases change. Regular reviews help you identify new gaps and check that agreed improvements remain effective.
04 / Supplier AI assurance statement
An independent statement, based on verified evidence, explaining how your organisation uses AI in delivering services and how well the assessed controls work. Its scope, evidence and limitations are explicit, helping buyers assess the assurance relevant to their tender or contract.
05 / AI policies and governance frameworks
We build the AI policies and governance frameworks you need, mapped to the requirements that apply to you and to the controls the assessment examined, so your approved use of AI is documented in a form you can evidence.
What we examine
- Who can access the tools, connected information and outputs.
- What information is exposed, shared or retained through the assessed use cases.
- Which permissions AI agents hold and what actions they can take.
- Whether impact and risk assessments address the actual use and its consequences.
- How accountability, human review and changes to approved use are evidenced.
The assessment is scoped to your controls and evidence. It does not certify a model’s accuracy or guarantee that every AI output is safe.
Applicable requirements
Requirements depend on who you are
01 / Victorian public sector
We use the Victorian Government’s generative AI guidance and OVIC’s enterprise AI guidance, together with the applicable protective data security and privacy requirements. We examine permitted use, information handling and the evidence supporting the organisation’s safeguards.
02 / Australian Government
We confirm whether the Policy for the responsible use of AI in government applies, together with relevant PSPF requirements and directions. The policy includes requirements for accountability, transparency, use-case registers, staff training and impact assessments.
03 / Government and Defence suppliers
We start with your contract, panel or buyer’s requirements and the information you handle. Agency policy does not automatically apply to every supplier in the same way. We assess the relevant expectations and use applicable official guidance to support the agreed scope.
04 / Organisations outside government
Where no government AI policy applies, we start with your privacy obligations, your contracts and the commitments you have made to customers. We use the Australian Government’s Guidance for AI Adoption and ISO/IEC 42001 as good-practice baselines, with ASD guidance informing the security assessment.
05 / A shared assurance foundation
The National framework for the assurance of artificial intelligence in government provides a common foundation across Australian governments. We use it alongside the requirements of the relevant jurisdiction, with ASD guidance informing the security assessment.
What you receive
For an AI adoption readiness assessment, you receive:
- An agreed scope statement naming the requirements, tools and use cases in scope.
- A readiness gap register against those requirements.
- Verification findings on access controls, information exposure and agent permissions.
- An independent review of impact and risk assessments for each use case.
- An independent readiness report with a prioritised roadmap.
Where selected, we also provide a supplier AI assurance statement, AI policies and governance frameworks, and regular assurance reviews.
Your organisation decides how to use AI and implements the safeguards. Sypha independently assesses them; we do not build AI solutions, configure tools, recommend vendors or provide legal advice.
What we need from you
A list of the AI tools and use cases in play, with approvals, impact or privacy impact assessments, your AI policy and training records, and evidence of who has access to what. Your team generates the access and identity evidence, and we verify it. We review existing material first and agree focused conversations with the people involved.
Official guidance
We confirm the requirements that apply to your organisation at the start of each engagement. These official resources provide the framework context; they do not imply endorsement of Sypha.
- Victorian Government: safe and responsible use of generative AI
- OVIC: enterprise generative AI tools in the Victorian public sector
- Digital Transformation Agency: Policy for the responsible use of AI in government
- National framework for the assurance of artificial intelligence in government
- ASD: AI data security
- Australian Government: Guidance for AI Adoption
- Protective Security Policy Framework: 2026 release
Read our guide to where to start with assurance when your people are already using AI.
A clear starting point
Discuss AI assurance needs
Tell us what you need to assess, the obligations you work under and the decisions ahead. We will agree a clear scope and the evidence needed to answer your questions.