Skip to main content
Aiviance
Legal centre

Aiviance AI Use Disclosure

Version 1.0 (AU) · Last updated 2 SEP 2026

Aiviance provides a software platform that helps individuals, families, businesses, and their advisors organise, analyse, and report on personal and business finances — via its website aiviance.ai, mobile applications, APIs and AI-assisted features (the “Service”).

The Service is operated by Aiviance Pty Ltd (ABN 13 688 678 980) ("Aiviance", "we", "us", "our"). Our current contact details are available on the Contact Us page.

This Disclosure describes how Aiviance uses artificial intelligence (AI) in the Service, including what data is processed by AI providers, your rights, and your options.

The Service helps you organise, analyse, and report on personal and business finances. The Service has some AI features, including:

  • embedded AI assistants;

  • AI-powered transaction categorisation and tagging;

  • AI-assisted drafting (notes, descriptions, summaries);

  • AI search and natural-language query;

  • AI-assisted analysis (cashflow, budget, anomaly, scenario);

  • AI document understanding (OCR, contract extraction, receipt parsing);

  • AI-assisted onboarding and conversational data entry.

When we say “AI”, we mean artificial intelligence systems and tools which may include machine learning techniques, large language models and other models that “learn”, which are embedded in the Service. We do not mean ordinary computational features (e.g. tax calculations from rule tables), statistical analysis without AI inference, or ordinary software features, algorithms or deterministic logic, or anything outside the Service.

“Customer Data” means any data, including financial records, transactions, documents, AI conversations, attachments, and other content, that you submit, generate, or receive through the Service.

Some bank data reaches the Service through the Consumer Data Right (CDR). This Disclosure covers AI processing of that data and of data derived from it. Because CDR data carries obligations of its own, two things apply to it that do not apply to other data. AI processing of CDR data takes place in Australia using approved processing infrastructure hosted in Australia. Withdrawing your CDR consent deletes the AI conversations that read data from that connection, together with anything the assistant remembered from them. The purposes for which CDR data is used are those set out in Fiskil's CDR Policy, which Aiviance has adopted as its governing policy as Fiskil's CDR Representative. The full position is in the CDR Consent Notice, and where this Disclosure and that Notice differ, the Notice prevails for CDR data.

1. AI Is Suggestive, Not Autonomous

Aiviance designs all AI features around a “review-first” principle:

  • AI proposes; you approve;

  • AI drafts; you confirm;

  • AI categorises; you can override;

  • AI analyses; you decide.

Importantly:

  • no AI feature autonomously posts a financial transaction, submits a tax return or BAS, initiates a payment, sends communication on your behalf, or modifies your tax position, super contributions, or investment positions;

  • no AI feature makes an automated or autonomous decision for you – it only makes suggestions (e.g. about data categorisations), all of which you can override;

  • any AI-made suggestions about financial state changes (transactions, journal entries, tax positions, payments) require explicit human confirmation before taking effect; and

  • AI-generated outputs that could influence a financial or tax decision are clearly labelled as AI-generated, with a recommendation to review.

2. AI Providers — Where Your Data Goes

Aiviance routes AI workloads to a curated set of third-party providers. We can review and change our service providers over time. Routing depends on your region and the data classification involved.

For more details on our current list of providers and routing arrangements, and their security controls and certifications, please refer to the sub-processor list at the Aiviance legal centre.

3. What Goes to AI Providers

Standard Rules

Aiviance classifies Customer Data into three sensitivity tiers as described in our Product Documentation available at the Aiviance legal centre. The Customer Data classification determines AI processing rules as follows:

  • Classification P0 — Highest sensitivity (tax file numbers; passport, driver licence and Medicare numbers; passwords and access tokens; card verification values; cryptographic keys; and full bank account and card numbers): This data is never sent to a generative AI or large language model provider. Some of it is blocked outright, so a request carrying it does not proceed; the rest is masked before transmission, leaving at most the last few characters. If you upload a document for text or field extraction, an approved OCR sub-processor may process P0 data contained in that document solely to perform that extraction. We design all other AI workflows to operate without P0 data — for example, by identifying an account by its nickname rather than its number. Where the Service enables you to type freely, please do not enter P0 data — we scan for it and remove it, but the reliable protection is not providing it.

  • Classification P1 — High sensitivity (transaction amounts, descriptions, counterparties and references; bank balances, account names and BSBs; merchant names; salary and deduction totals; your name and entity names; and your contact details and personal identifiers): Before a request is sent, your date of birth is replaced with a non-reversible value, your telephone number is masked, and your email address is partly masked. Everything else in this tier is sent as it appears, and only to an AI provider operating under a zero-data-retention agreement, under which your prompts and the AI's responses are not kept by that provider and are not used to train any model. A request to a provider without such an agreement is blocked before it leaves Aiviance.

  • Classification P2 — Standard sensitivity (account preferences; categorisation labels without the underlying data; region, currency and chart-of-accounts codes; and public-register identifiers such as an ABN or ACN): This data can be sent to AI as it is.

Classification Override

You can adjust classification preferences for the purposes of AI use in Profile Settings → Privacy → Data Classification. You can choose not to permit P1-classified data to be used with AI at all. If you do, a message containing such data is not sent — the Service tells you so and offers to continue without those details — and some AI features are disabled altogether. You can also elect to restore to the default settings described above.

4. How We Use AI

  • Categorisation and tagging — AI can suggest categories for transactions, expenses, and income; you can review, edit, or reject suggestions and set your own categorisation rules.

  • Drafting and summarisation — AI can help draft notes, descriptions, summaries, and explanations; relevant content is labelled as AI-generated and can be edited by you.

  • Search and natural-language query — AI can use natural-language instructions to find financial data that you are authorised to access. Search itself only reads data; it does not modify records.

  • Analysis and insights — AI can generate cashflow analysis, budget comparisons, anomaly flags, and scenario projections. These outputs are for information only and do not constitute financial product, tax, or legal advice.

  • Document understanding — where the relevant feature is available, AI can extract text and fields from receipts, invoices, statements, and other documents that you upload. You must confirm any result that may update a financial record.

  • Conversational data entry — AI can organise transaction information that you provide in natural language into a structured draft. You must confirm the draft before it is saved or posted.

Some features that look similar are not AI-driven, e.g.:

  • tax calculations performed under established rules;

  • currency conversion using external exchange-rate data;

  • aggregate sums and totals (deterministic); and

  • compliance checks against fixed rule tables.

5. Risks and Limitations in AI Outputs

AI outputs can be wrong or misleading. In particular, AI outputs might:

  • repeat or perpetuate errors in Customer Data or prompts (because they are generated based on Customer Data and prompts);

  • contain errors, for example, where the AI has misinterpreted an input or Customer Data;

  • make omissions;

  • “hallucinate” (i.e. create plausible-looking content that is actually incorrect);

  • rely on model knowledge, data sources, or context that may not be current, and therefore fail to reflect recent regulatory changes, practices, market conditions, or product features;

  • miscategorise transactions or misread what you intended; and

  • reflect biases inherent in the data the AI model was trained on.

Aiviance uses technical controls, rule-based checks, and quality evaluations to reduce the risk of certain errors, unsupported claims, and inconsistent outputs. These measures do not guarantee that AI outputs will always be accurate, complete, or free from bias. It is therefore essential that you:

  • verify your inputs and Customer Data for accuracy and completeness;

  • always review and verify AI outputs before accepting or relying on them, and use available features to confirm, edit or reject AI outputs;

  • check outputs against authoritative, up-to-date sources especially where there have been recent regulatory or market changes; and

  • periodically review and correct categorisation results and, where available, set categorisation rules to improve the consistency of future suggestions.

If you have concerns about an AI output such as a biased output, please report it to support@aiviance.ai.

AI outputs do not constitute financial product, tax, or legal advice, and do not take account of your personal objectives, financial situation, or needs.

6. AI Conversation History and Retention

Aiviance retains AI conversations for the period you select, so that it can provide context across sessions and let you review past conversations and AI outputs. De-identified, aggregated usage statistics may be retained separately to improve service quality, evaluate feature use, and investigate problems.

The default retention period for AI conversations is 90 days. In Profile Settings → Privacy → AI Conversation Retention, you can select an available retention period from a minimum of 7 days to a maximum of 365 days. A shorter retention period may limit features that rely on context from earlier conversations.

You can also delete all AI conversation history at any time in the Service. A deletion request removes the relevant conversations from Aiviance's active systems within 24 hours. Copies in backups expire through the rolling backup cycle within 90 days, unless Aiviance is required by law to retain them or they are subject to a legal hold.

7. Data Not Used to Train AI Models

Aiviance does not use Customer Data to train or fine-tune Aiviance AI models, general-purpose AI models, or third-party models, and does not permit AI providers to use Customer Data for those purposes. Aiviance sends Customer Data only to approved providers for which model-training use has been excluded.

Aiviance may use aggregated, de-identified usage statistics that do not identify individual customers to improve its products and services. These statistics are not used to train or fine-tune AI models.

8. Automated Decision-Making

Aiviance does not use AI or any other tools to make solely automated decisions that produce legal or similarly significant effects on you.

AI-generated categorisations, drafts, analyses, and other suggestions do not take effect on their own. Any AI suggestion that may modify a financial record, initiate a payment, submit a filing, or produce another external effect must be confirmed by a person in accordance with the applicable permissions and approval rules.

If Aiviance introduces a feature in future that may constitute automated decision-making, Aiviance will update the relevant disclosures before the feature is enabled and provide any notice, human review, objection, or opt-out mechanism required by applicable law.

9. Your Rights and Controls

You have the following rights and controls:

  • Opt out of AI features: In Profile Settings → Privacy → AI Features, you can disable AI features in the Service. You retain access to non-AI features.

  • Adjust data classification: In Profile Settings → Privacy → Data Classification, you can prevent P1-classified data from being used with AI features.

  • Review AI outputs: You can confirm, edit, or reject any AI suggestion that may modify a financial record or produce an external effect before it takes effect.

  • Delete AI conversation history: In Profile Settings → Privacy → AI Conversation History, you can delete all AI conversations.

  • View AI processing records: In Profile Settings → Privacy → AI Routing, you can review recent AI processing records, including the provider, processing jurisdiction, data classifications, and relevant safeguards.

10. Updates

This Disclosure may be updated over time to reflect changes in AI providers, models, routing, data classification rules, processing practices, or regulatory requirements.

For material updates affecting the scope of AI processing, external providers, or data retention practices, Aiviance will notify users by email and in-product notice at least 30 days before the update takes effect.

An update required by law, or reasonably necessary to address a security, fraud, or other urgent risk, may take effect on shorter notice or immediately. Where reasonably practicable, Aiviance will notify you as soon as possible.

Clarifications, minor corrections, and model version updates that do not materially change how Customer Data is processed take effect when the updated Disclosure is published.

11. AI and Children

The Service is not for direct use by children under 18. Parents, legal guardians, and other adults with legal authority may use the Service to record child-related expenses and childcare or education information.

Child-related data entered through those features may be processed by the AI features described in this Disclosure and may be visible to authorised members of the relevant family workspace. The adult entering that data is responsible for ensuring that they have authority to provide and manage it.

12. Contact

For questions or complaints about AI use, this Disclosure, or related data processing, email the Aiviance Privacy Officer at privacy@aiviance.ai.

To contact us by post, use the mailing address listed on our Contact Us page.