How automated trading works in Australia, from alerts to execution

Automated trading in Australia covers a far wider range of things than most people expect. At one end it is an alert telling you a condition has been met. At the other it is code that sizes a position, sends the order and manages the exit while you sleep.

"Algo trading" is the same idea described from the institutional side. ASIC has put algorithmic trading at roughly 85 per cent of turnover in Australian listed equities, so anyone trading is already trading alongside algorithms.

This is general information, not personal advice.

What automated trading actually means

Strip the marketing away and automated trading is two things bolted together: a set of rules, and a mechanism for acting on them. The rules might be one price level or a dozen inputs; the mechanism might be your hand on a mouse, a script on a schedule, or a live connection into a broker's order system.

Keeping those apart matters, because they fail differently. A sound rule executed badly bleeds through slippage and missed fills, and an unsound rule executed perfectly loses money faster than you would by hand. Automation removes hesitation and does nothing to make a signal predictive, which is why the piece on whether AI can predict stock prices belongs beside this one.

The spectrum, from alerts to unattended execution

The first rung is alerts. Software watches for a condition and tells you when it is met, and you still place the order. Start here, because it forces you to write the condition down. There is a practical guide to setting alerts you will actually read.

The second is rules you follow mechanically by hand: entry, exit, position size and when to do nothing. The third is the order automation your broker already provides, such as stop orders, trailing stops and linked orders where one cancels the other. Those address the most common failure in retail trading, not closing a position you said you would close.

The fourth is unattended execution, your own code deciding and sending orders with nobody at the desk. That is a software operations job as much as a trading one, needing clean data, error handling, reconciliation and a fast way to stop everything. For an Australian trading US listed assets that session runs overnight, so a system trading while you sleep is the attraction and the risk at once. Most retail investors get most of the benefit from the first three rungs.

What brokers in Australia allow, and what they do not

Programmatic access is the broker's decision, not your entitlement, because ASIC's electronic trading guidance makes the market participant the gatekeeper for everything reaching the market through it. The participant has to keep direct control of its automated filters, be able to suspend all automated order processing immediately, cancel orders already in the market, and keep seven years of trading records. ASIC is explicit that arrangements bypassing those controls are not permitted, and that a client should first have demonstrated knowledge of the broker's order entry system, which is why API access comes with agreements and eligibility conditions rather than a simple toggle.

In practice, plain share broking accounts often offer no programmatic access, some provide data without order entry, and a smaller group offer order APIs with those conditions attached. Platforms for leveraged derivatives and forex usually ship with a scripting language and automated execution built in, the easiest route to a running system and where sizing mistakes do most damage.

Before going offshore for a friendlier API, note Moneysmart's position. Anyone offering investments here must hold an Australian financial services licence or be authorised by one, and international investments not offered by licensed Australian providers sit outside ASIC's regulation, so you are not protected if something goes wrong. See also the overview of AI investing tools in Australia.

Where automated systems come apart

Overfitting is the big one. Given enough parameters and enough history you can fit rules to anything, and the result describes noise. Two quieter versions do the same damage: survivorship bias, where the universe you tested holds only the names that survived, and look ahead bias, where a signal uses information that was not yet public.

Then there is the gap between a simulated fill and a real one. Backtests fill at a convenient price, instantly and in full, while real orders cross a spread, queue, fill partially and sometimes gap straight past your exit level. Brokerage and currency conversion take their share too, and a thin edge rarely survives that.

Data quality failures are the hardest to spot, because the system keeps running. Splits, ticker changes, delisted names and erroneous prints corrupt a signal quietly, and time zones worst of all, since Australia sits ten or eleven hours ahead of UTC.

Last, stop discipline. Most people want automation to stop them overriding their own rules, then keep a manual override anyway, and a system with an override is a discretionary system with extra steps. You also want your own version of what ASIC requires of participants: a limit that halts the whole system, not one trade.

What the regulator expects of automated trading in Australia

Most of the rulebook is aimed at brokers and market participants, not at individuals trading their own money. Trading your own account, with rules or software, is not itself a licensed activity.

The parts applying to everyone sit in the Corporations Act 2001, which prohibits market manipulation under section 1041A and false trading and market rigging under section 1041B. Those prohibitions attach to the conduct rather than the tool, so they bite whether a person or a program sent the orders.

ASIC has also moved on this specifically, amending the market integrity rules for securities and futures markets so trading system obligations cover algorithmic trading, artificial intelligence and machine learning, with the same requirements applying to manual and automated orders. It calls the approach technology neutral, and the amendments take effect in 2028.

The line individuals do cross is sharing a system with other people. ASIC's position is that anyone carrying on a business of providing financial services needs a licence unless an exemption applies, and that arranging for someone else to deal can itself be a financial service. Running a bot for friends for a cut is worth proper advice first.

How InsiderPulse handles this

InsiderPulse sits at the data and scoring end of this spectrum, not the execution end. It combines SEC Form 4 insider filings, US congressional trade disclosures, options activity, dark pool and volume data, prices and news into one score from 0 to 100, covering US listed stocks and ETFs plus crypto, forex and commodities. It does not place trades and does not connect to a broker, so you read the data and decide what to do in your own account. The page on how the score is built sets out the inputs.

InsiderPulse is a data tool. Nothing on this page is financial advice.