Traditional financial services business structures are under threat from rapid digital innovation. Financial advisers are increasingly being replaced by sophisticated computer programs.

The capability of robo-advice has quickly evolved from simple website calculators to software and websites providing financial advice, portfolio construction and active investment management. It is one of the fastest growing industries in the United States and United Kingdom.

In Australia, Stockspot, InvestSMART, Ignition Wealth, Yellow Brick Road, NAB and most recently Macquarie have all embraced robo-advice. However, deploying robo-advice is not without issues. The current legal framework in particular is a minefield – the result of laws not keeping up with technology.

Here we outline the legal impediments and potential solutions to deploying robo-advice in your organisation.


There are myriad potential regulatory consequences when developing roboadvice capabilities. As a general guide, the greater the functionality, the greater the compliance challenge.

For example:

  • providing “factual advice” does not require an AFSL;
  • providing “general advice” requires an AFSL (or an exemption) and a “general advice warning” to be given to the client; and
  • providing “personal advice” requires an AFSL (or an exemption) and a statement of advice to be given to the client. In giving the advice, the “best interests” duty and associated obligations under the FOFA regime will apply.

Before forging ahead and incurring development expenses, it is worth understanding the regulatory implications – the incremental functionality may not be worth the substantial increase in compliance costs.


Banks and financial institutions considering deploying robo-advice are highly attracted to the possibility of utilising the wealth of personal information they have about their customers to provide extremely tailored and relevant advice.

For instance, existing information about a customer’s age, marital status, number of dependants, occupation, salary, superannuation, savings, borrowings, risk tolerance and future objectives can all be used to build a detailed financial picture of a robo-advice client. This information can be used, in turn, to generate investment strategies and product recommendations.

The problem is that using personal information comes with a high risk of falling into the realm of “personal advice” – bringing with it very stringent duties and disclosure obligations under the Corporations Act. There is no clear cut test to determine when using personal circumstances takes advice into the realm of personal advice.

As a guide, personal circumstances can be used to give general advice that is more relevant to the client. However, it is crucial that any general advice given is not presented in a way that means a reasonable person would expect one or more of the client’s objectives, financial situation or needs to have been considered. If those things are considered, the advice is likely to be “personal advice”.


Providing only “factual advice” may seem like an attractive option to avoid the need for an AFSL and disclosure to potential clients. However, careful attention needs to be paid to the messaging around factual advice to ensure it does not fall into the realms of general or personal advice.

“Factual advice” is objectively ascertainable information about financial products, the truth or accuracy of which cannot reasonably be questioned.

The advice must not be presented in a way that is intended to, or can reasonably suggest or imply an intention to, make a recommendation about what a client should do. Accordingly, care must be taken to ensure that no qualitative judgement, evaluation or assessment of the features of the products is presented alongside factual information given by the robo-adviser.

In our experience, developing marketing messages to sell the product is a particularly difficult area because these messages tend to add a value-judgement about the service itself to what is otherwise factual information (for instance, “our product only selects the best stocks in the market”).

Those with robo-advice capability therefore need to look holistically at the information they offer, in the context of how their robo-advice service is represented, to avoid their product offering inadvertently suggesting that a client should take a particular course of action or purchase a particular product.

As is the case with general advice, personal circumstances can still be used to tailor the factual information presented. However, utilising personal circumstances to deliver more relevant factual advice to the client can suggest that the output presented is something more than just “factual advice”.

It is essential that, after being refined according to the client’s personal circumstances, the information presented remains objectively ascertainable and offers no recommendation or statement of opinion intended to influence the client – otherwise the robo-adviser may actually be giving “personal advice”.


Robo-advisers giving personal advice must comply with the FOFA regime – in particular, the requirement to act in the “best interests” of the client and to give “appropriate” advice.

Satisfaction of the best interests duty is measured by reference to a “safe harbour” test, which lists a number of matters that must have been considered by the adviser. There are serious questions as to whether robo-advisers are able to meet the best interests safe harbour tests.

These concerns stem from three realities about the current state of computer technology:

  1. The architecture of the program delivering the robo-advice can only produce a series of pre-defined outcomes. Until computers are able to truly “learn” and independently investigate, all outcomes are determined prior to the client receiving any advice.
    Accordingly, applying a predefined set of criteria to each client’s unique personal circumstances runs the risk of the client not receiving appropriate advice. This risk may eventuate by the program failing to consider all of the client’s relevant circumstances in preparing the advice (potentially not satisfying section 961B(2)(f) Corporations Act) or as the roboadvisers’ universe of financial products is pre-determined, by failing to make a reasonable investigation into the financial products that achieve the objectives and needs of the client (potentially not satisfying section 961B(2)(e) Corporations Act).
    In addition, as a result of the pre-defined range of possible outcomes, there is a possibility of, intentionally or unintentionally, programming a conflict of interest into the robo-advice.
    This could take the form of, for example, the programs in-built “approved product list” only including products issued by related entity. As the program would be unable to recognise this when giving personal advice, the advice may fail to respect the duty to prefer the interests of the client in the case of a conflict of interest. Accordingly, there is a risk of breaching section 961J Corporations Act.
  2. Robo-advisers may not be able to recognise when information is incomplete or inaccurate – especially when that information is being provided by the client.
    This presents a risk of not satisfying section 961B(2)(c) Corporations Act which requires that reasonable investigations are made to obtain complete and accurate information when it is reasonably apparent that the information relating to the client’s circumstances is incomplete or inaccurate. “Garbage in, garbage out” is a phrase to remember when considering 961B(2)(c) issues – the robo-advice is only as good as the information inputted by the client.
  3. The best interests safe harbour contains a catch-all in section 961B(2)(g) Corporations Act.
    It requires that “any other step” is taken at the time the advice is given if that would be in the best interests of the client given their circumstances. Where a determinative process is followed as a result of the robo-advice program, there is a risk that section 961B(2)(g) may not be satisfied.


Robo-advice is still in its infancy in Australia and various regulatory issues are yet to be completely addressed. For example:

  • Section 912A(1)(f) Corporations Act requires a financial services licensee to ensure that “its representatives are adequately trained, and competent, to provide those financial services”. Where the adviser is a computer program, who is the appropriate person to be adequately trained and competent? Does the programmer of the robo-adviser need to meet these standards? Are regular updates (“computer continuing professional development”) required to ensure the robo-adviser remains up to date with the latest trends and newest financial products?
  • If you have a dispute with your robo-adviser, how would this be resolved? What compensation arrangements would be available to you?


ASIC is supportive of the automatic provision of advice for its possibility to offer a convenient, low-cost advice service to consumers. In addition, ASIC sees benefits such as improved compliance and record keeping and the potential to reduce conflicts of interest.

ASIC is closely monitoring the development of robo-advice in Australia and is engaging with robo-advice providers operating in Australia. ASIC views the obligations imposed by our financial services laws as “technology neutral” and has observed that many start-up robo-advisers struggle with the fine distinctions between factual, general and personal advice.

While ASIC has not published any formal guidance, it does recognise the difficulties with our existing legal framework and has established a “robo-advice taskforce” to investigate:

  • how robo-advice providers comply with the best interests duty;
  • how robo-advice providers develop and test their algorithms;
  • the training and competency requirements for those sitting behind robo-advice models; and
  • the adequacy of a robo-advice operator's compensation arrangements.

Although no reporting date has been set for the findings of the taskforce, many in the industry are eagerly awaiting the taskforce’s recommendations and further guidance from ASIC.

Corrs’ comment

The significantly lower cost, sizeable market of Australians not currently receiving any financial advice (especially the SMSF sector) makes robo-advice a highly attractive proposition. Robo-advice programs also have come with the assurance that mandatory disclosures are complete and always given at the right time – a distinct advantage over human advisers.

However, there are legal obstacles that need careful management and bespoke legal advice prior to launching.

While it’s unlikely the FOFA regime can again be refined to address the robo-advice legal issues, it’s clear the advice industry in Australia is getting on board with computer-generated advice.

ASIC’s clear enthusiasm for the possibilities of robo-advice should give confidence to the industry that ASIC will appropriately balance product innovation and regulatory intervention.