Speaking to SMSF Adviser, Super IQ’s chief executive Andrew Bloore said a fully automated audit process is something Super IQ’s auditors are looking to achieve.
“Our auditors are working with us about designing appropriate ways of doing this,” said Mr Bloore.
“Within a period of two years I strongly believe we will have an integrated audit process, and much better tools to deliver information to trustees,” he said.
Improved auditing, Mr Bloore said, will mean trustees no longer need to inform the administrator, auditor and the tax office if an event has occurred because it will be automatically picked up.
“We’ll get it the instant the trustee does it and it will flow right through the whole system so at any point in time everyone is completely informed,” he said.
Smithink’s founding director David Smith on the other hand expects the time frame for fully automated auditing will be closer to 10 or 15 years.
The ability for the industry to make the audit process fully automated, he said, will depend on administration platforms having the “energy and desire to open themselves up to provide the [required] information”.
Auditors, he added, will also need to build highly intelligent software engines that can search for all the bits of information needed to ensure everything is done appropriately – “I don’t think that’s going to happen tomorrow”.
The data from the administration systems will also need to be certified under the GS007 standard for auditors to be able to rely on the systems and automate the process, Mr Smith said.
“At this time none of the administration platforms have done that – the reason they haven’t done that is because it’s pretty expensive to do so,” he said.
Mr Smith said in any audit there is also an element of judgment with auditors making judgments about valuations of assets for example, and how assets are disclosed in financial statements.
“Those things are at the moment pretty much subject to human judgement,” he said.
The question he said is how far away machines are from making those judgements.
“My opinion at the moment is probably a 10- or 15-year scenario, if not longer,” he said.
“We’re talking about automating something which is quite a complex judgement decision; that said machine automation is moving at a cracking pace so maybe it’ll be quicker than I think.”
In the nearer term Mr Smith said the SMSF industry can expect to see a lot of the manual procedural work in auditing stripped away.
“I think we’re going to be seeing that happen in a fairly rapid rate of knots because I think some of the self-managed super auditing companies are making a fair amount of investment, and auditing platforms, like Auditflow, are building more and more automation in the audit process,” he said.
Lord
What Andrew and David in this article seem to think that a software can be developed where a machine can “pass judgement” on the validity of each and every transaction.
My comment is “Yes” provided the machine is told every parameter. Here is another example: Section 66, acquiring an asset from a related party – the machine has to be fed with name of every related party including part 8 associates of each member of the fund for that function to work.
By the time we data feed each and every related party for each year, the human time will be more than human time to actually “pass that judgement” that the asset has or has not been acquired from a related party.
When auditing a fund – we can and have built an audit tool, which can audit 90% to 95% of all financial transactions, but 0% on compliance. It has some business intelligence.
Manoj,
Using your related party loan scenario, I am sure a software house (like your own) can write a sub routine that asks 5 or 6 questions regarding the loan. Based on those answers, determine a logical outcome or at a minimum, redflag it for further consideration.
Australian auditing standards requires auditors to independently verify all transactions. But a SMSF audit is not only about a financial audit, it is more to do with compliance and most contravention are relating to compliance. No auditor complains that balance sheet does not balance.
For example, say some money has been lent by a SMSF, a machine can correctly classify that it is a loan from the SMSF, but the machine has also to be trained to decide that the receiving party is not a related party. To do this, you will have to provide details of all relatives (Grandfather, uncles and nieces etc) of all the four members of the fund and detail all the companies and trusts where relatives have control who is a partner of each member every year.
In my opinion automatic audit is only a dream wish. Financial audit can happen.