AI for financial advisers in Australia.

AI for financial advisers is software that prepares advice work: drafting file notes, organising fact-find information, assembling SoA request packs, filling paperwork and preparing reviews. In Australia it does not change who is responsible. The adviser still decides the advice and acts in the client’s best interests, and the licensee still answers for how AI is used.

ASIC has said the financial services rules are technology neutral, and that licensees should not wait for AI-specific reform before applying them (ASIC REP 798).

The useful work is the prework between a client meeting and advice that is ready to present. Our guide to the financial advice process sets out each stage. These are the places AI most often helps.

Worked example

Illustrative example with fictional names. An AI tool drafts a discovery file note for adviser Alex Example. The draft says the client “will retire in 2027”. The client actually said they might cut back to three days a week. Alex corrects the note before it is saved.

The draft spared Alex the typing. It did not replace Alex’s check.

Some decisions belong to a person, whatever software you use:

• Advice judgement. Strategy and product recommendations, and whether advice is needed at all.
• The best interests duty. The Corporations Act requires the advice provider to act in the client’s best interests (s 961B), give appropriate advice (s 961G) and prioritise the client’s interests (s 961J). ASIC explains these duties in RG 175. A tool cannot carry them for you.
• Approvals. Client authority, adviser sign-off and authorising each implementation action are separate decisions. Each should be made by a named person and recorded.
• Client consent. Agreement to record a meeting, ongoing fee consent and anything else that needs the client’s agreement stays explicit.

In ASIC’s review, AI generally augmented rather than replaced human decision making. A clickthrough is not advice approval: if a reviewer approves a draft without reading it, the check has not happened.

ASIC’s main statement is REP 798 Beware the gap: Governance arrangements in the face of AI innovation, published on 29 October 2024. ASIC analysed 624 AI use cases that 23 licensees in banking, credit, insurance and financial advice were using or developing as at December 2023. The findings that matter most for advice firms:

• Adoption is speeding up. 61% of licensees in the review planned to increase their AI use in the next 12 months.
• Governance did not always keep up. Some licensees were updating governance at the same time as increasing AI use, and for two of them governance lagged behind it.
• Policies had gaps. Only 12 licensees had AI policies that referenced fairness, and only 10 referenced disclosing AI use to affected consumers.
• Third parties need governing too. Many licensees relied heavily on third parties for AI models, and not all had arrangements in place to manage the risks.

ASIC named the obligations to consider first: the general licensee obligations, directors’ duties and the consumer protection provisions, including the bans on unconscionable conduct and false or misleading representations. The report ends with 11 questions for licensees, including who is accountable for AI and whether governance is leading or lagging AI use.

What ASIC has said since

ASIC’s Corporate Plan 2026–27, published on 26 August 2026, lists the impact of AI use among its focus areas. On 8 May 2026, ASIC wrote an open letter to AFS licensees on frontier AI and cyber resilience, asking them to review user access, manage third-party risk and strengthen basic controls. And on 23 March 2026, ASIC’s Moneysmart published consumer tips on using AI for money questions, so expect more clients to arrive with AI-generated ideas to talk through.

The main reference is the OAIC’s Guidance on privacy and the use of commercially available AI products, published on 21 October 2024 and updated on 17 January 2025. Its key points for advice firms:

• Privacy obligations apply to personal information put into an AI tool, and to output that contains personal information.
• Do due diligence before adopting a product, then keep reviewing it. The OAIC warns against a “set and forget” approach.
• Take a privacy-by-design approach, including a privacy impact assessment.
• Update your privacy policy and notices to explain how you use AI.
• As best practice, do not enter personal information, particularly sensitive information, into publicly available generative AI tools.
• If AI generates or infers personal information, that is a collection under APP 3. The guidance uses AI meeting minutes as an example.

A new rule starts on 10 December 2026. The Privacy and Other Legislation Amendment Act 2024 added APP 1.7 to 1.9. From that date, privacy policies must include information where the firm has arranged for a computer program to make or inform decisions that significantly affect individuals’ rights or interests. The OAIC published guidance on 30 September 2026. Whether any of your tools are caught is worth asking your licensee or privacy adviser now.

Some small businesses are exempt from the Privacy Act. If you are not sure whether your firm is covered, check with your licensee. The OAIC guidance is a sound standard either way.

Shadow AI is staff using AI tools for work outside an official policy or approved approach. That is the National AI Centre’s definition, and it notes that staff are likely already using AI at work, whether or not the organisation has a policy.

In an advice firm it usually looks ordinary. Illustrative example: Jordan Example, a fictional client services officer, pastes a client’s super statement and tax return into a free chatbot to summarise them before a review. It is quick and well meant. But the firm cannot now say where that data went, who can see it, whether it will train a model or how long it will be kept, and nothing is on the client file. It is exactly what the OAIC recommends against.

A ban rarely works on its own. The National AI Centre suggests making it safe for staff to say what they use, giving clear guidance on which tools are acceptable, and investing in training and policies. For an advice firm, that means:

• An approved list. Which tools are allowed, and what each may be used for.
• A governed alternative. A tool that does the job people were reaching for, under the firm’s control and permissions, with a record of use.
• A rule for client data. No personal information goes into public AI tools.
• A way to ask. Staff can check before trying something new.

Ask these of any AI tool, including AI features added to software you already use. Get the answers in writing, and record who approved the tool.

The government’s framework for the answers

The Voluntary AI Safety Standard, published on 5 September 2024, set out 10 guardrails. On 21 October 2025, the National AI Centre published the Guidance for AI Adoption, which updates and simplifies the standard into six essential practices: decide who is accountable, understand impacts and plan accordingly, measure and manage risks, share essential information, test and monitor, and maintain human control.

Both are voluntary. Neither is a certification, and no tool meets them on your behalf: they describe how your firm governs its own use of AI.

Most tools fall into four groups, and many span more than one. The checklist applies to all of them.

This guide does not rank or recommend tools. Choose on the answers to the checklist, tested against your own files.

This guide is general information, not legal or compliance advice. Confirm how these rules apply to your firm with your licensee or compliance adviser.

Bunya AI builds Bunya Command Centre, the practice operating system for Australian financial advice firms. It is built and installed in your firm’s own Microsoft 365, and owned by your firm. AI workflows prepare file notes, fact-find inputs, SoA request packs and review preparation, and a person on your team checks and approves every step before anything moves on. AI-prepared work is labelled, and AI usage can be recorded with the client record.

The AI services used, where data is processed and their data-use terms are reviewed with your firm as part of the solution design. Bring the checklist above to that conversation.

• AI workflows & integrations: what each workflow uses, what it prepares and who checks it.
• Security & your data: your tenancy, access control, audit trail and ownership.
• Bunya Voice: an optional add-on for calls, meetings and follow-up, including file notes.