Aura8 AI. With its act together. Free Risk Check
Governance Decision Pack

Answer your customer's AI due diligence with proof behind every answer.

A fixed-price, evidence-graded review for AI vendors selling into banks, insurers and super funds. We grade up to 20 questions from your customer's AI or security due diligence questionnaire against your own records, so nothing you send is a claim you can't back.

7 business days
From receipt of your evidence pack
Up to 20
Substantive AI questions answered and graded
AUD 4,950 + GST
Fixed price

Lender, BNPL or lending-tech firm getting ready for 10 December? See the AI Customer Assurance Sprint. AI vendor answering a bank, insurer or super fund's due diligence questions? See the Governance Decision Pack.

Where deals stall

A bank, insurer or super fund has sent you a due diligence questionnaire, and somewhere in it is a section on AI. It asks how you control changes to your models, where your training data came from, which foundation model sits behind your product, whether a person can step in on high-risk decisions, and whether the institution's data is ever used to train anything.

Your team does most of these things. The hard part is proving it. Your CTO stops building and starts drafting from memory. The answers read well, but nobody can say which ones your records actually support. Each weak answer comes back as another round of questions, and each round moves the deal date.

Institutions are looking harder at the AI and the third and fourth parties behind their suppliers, and APRA-regulated entities now manage their service providers under CPS 230. Answers get tested, not just read.

How it works: Map, Grade, Prove

Map

We separate the AI questions from security, legal and procurement, map each one to the record that could prove it, and map the AI supply chain behind your product: model providers, hosting, subprocessors and fourth parties.

Grade

We grade every proposed answer against your evidence.

Prove

We write each response to its grade, build a dated evidence register and a ranked gap list, and hand it all over in a 60-minute decision session.

The four grades

GradeWhat it meansWhat we do with it
SupportedA dated record backs the answer as writtenSubmitted as written, with the record cited in the register
Partially supportedA record exists but is incomplete, undated, or covers only part of the claimThe answer is narrowed to what the record supports, and the gap goes on the priority list
Not evidencedNo record backs the claim, even if the practice existsFixed before submission, or disclosed with a plan. Never submitted as if it were Supported
Not applicableThe question doesn't apply to your productAnswered as not applicable, with the reason stated

The grades are evidence-readiness judgements to support your decisions. They aren't legal advice, a certification or an audit opinion.

What's included

Submission-ready, evidence-graded AI responses

Up to 20 substantive questions, each with its grade, the answer written to the grade, and the record it relies on.

AI supply chain and fourth-party map

Model providers, hosting regions, subprocessors and fourth-party dependencies, with concentration flags.

Dated AI evidence register

Every record with its owner, date and next review date, so the next questionnaire starts from evidence, not memory.

Evidence-gap priority list

Each gap ranked (fix before submission, disclose with a plan, or no action), with suggested disclosure wording.

A 60-minute decision and handover session with your decision owner

We walk through the grades and the gaps, and you decide what goes to the institution.

You decide what goes to the institution. We make sure every answer you send is one your records support.

Timing and what we need from you

WhenWhat happens
KickoffEngagement signed. We send the evidence request and a short pre-questionnaire for your technical owner (about 30 minutes).
Clock startsThe 7-business-day clock starts when your evidence pack arrives. You commit to sending it within 2 business days of kickoff.
Days 1 to 2Map the AI questions to your records and your AI supply chain. One 90-minute interview with your technical owner.
Days 3 to 4Grade every answer and complete the supply chain map. A 30-minute clarification if needed.
Days 5 to 6Draft the responses, rank the gaps, build the register.
Day 7Pack delivered, with the 60-minute decision and handover session.

From you: a named evidence owner who pulls the records, your evidence pack within 2 business days of kickoff, your technical owner for no more than 3 hours in total, and your decision owner at the Day 7 session.

Price

AUD 4,950 + GST, fixed price. It covers up to 20 substantive AI questions. Larger questionnaires are quoted before we start.

Not included: legal advice, certification, an ISO 42001 implementation, SOC 2 or ISO 27001 remediation, penetration testing, contract negotiation, a full cybersecurity assessment, or any promise that the institution will approve you or sign.

What we won't promise

That the institution approves you or signs. Nobody outside the institution controls that. What the Pack does is stop the AI section being the reason your deal slips.

See it before you buy

A sample Pack for a fictional supplier, marked SAMPLE on every page: graded answers, a supply chain map and a register excerpt.

How it fits the Guidance for AI Adoption

The Australian Government's Guidance for AI Adoption (October 2025) sets out six essential practices for organisations using AI, including deciding who is accountable, measuring and managing risks, and sharing essential information. Your customers are applying the same thinking to their suppliers. The dated register and supply chain map give you a clear record of where you stand against those questions.

FAQs

We have SOC 2 and a compliance library. Can't we do this ourselves?
You can draft it yourselves. SOC 2 shows your security controls operated, but it wasn't scoped for model change control, training data provenance, foundation-model dependencies or human oversight. A library stores answers. It doesn't tell you whether each one is still true for the model you shipped last month. Grading and dating do that.
What does it cost, and what if we have more than 20 AI questions?
AUD 4,950 + GST, fixed price, for up to 20 substantive AI questions. Large institutions often send more. We confirm the count before we start and quote any extra questions then, so the price is agreed before work begins.
When does the 7-day clock start?
When your evidence pack arrives. You commit to sending it within 2 business days of kickoff, and delivery is 7 business days from receipt.
What if you find gaps? Do we have to tell the bank?
You decide what goes in. Our rule is simple: we never write an answer we graded Not evidenced as if it were Supported. A gap disclosed with an owner and a date is a conversation about a plan. An overclaim the institution finds later is a credibility problem.
Is this legal advice or a certification? Will the bank approve us?
No to both. The grades are evidence-readiness judgements. We don't give legal advice, certify, or issue audit opinions, and we can't promise the institution's decision. Where you need a lawyer, we'll say so.

Lender, BNPL or lending-tech firm getting ready for 10 December? See the AI Customer Assurance Sprint. AI vendor answering a bank, insurer or super fund's due diligence questions? See the Governance Decision Pack.

Book a free Risk CheckFree Risk Check