Auditing as it is currently practised is already on the way out. Algorithm based tools such as MindBridge and in-house algorithms developed by the largest firms are turning reconciliation, error checking and flagging suspicious transactions into a push-button affair.
Auditors will still be around – but they could look and sound a lot more like Jules and Vincent from the classic gangster flick Pulp Fiction than Louis “come in for a low-sodium mineral water” Tully in Ghostbusters. Or perhaps a TV series, F-CSI – Financial Crime Scene Investigation?
Few are as heavily involved in this transition than chief data scientist at Deloitte Australia, Gavin Whyte. Whyte spent years developing algorithms in anti-money laundering, marketing optimisation and transactional processing for gambling companies and banks before joining the Big Four.
Last year he wrote an algorithm that automates bank reconciliation and can process hundreds of transactions in seconds. In a Digital First webinar last week (you can watch it at the end of this article), Whyte says the impact of these algorithms will force the role of auditors to change from back-office bandits to an outsourced investigation unit.
“I’ve literally said this several times. Deloitte did a report last year on the number of jobs that is going to be automated by AI. They found that auditors were right on top of the list.
“Auditors will eventually veer towards the forensic accounting, accuracy, validation type of role rather than sitting with Excel spreadsheets trying to manually reconcile thousands of transactions,” says Whyte.
These machines can already self-learn – they literally program themselves to recognise suspicious transactions or allocate tax codes. This is great news for SMEs but not so great for accounting firms’ revenues.
“Forensics charges or accounting charges will drop purely because they do things quicker and faster, producing far more accurate results and better serving the client at the end of the day,” Whyte adds.
If someone claimed $90 for a cab ride from the city to the airport when the average cab charge for that distance is $50, the algorithm will flag the transaction as an anomaly, Whyte says. Instead of spending weeks and months finding transactions to investigate, auditors could instantly get a list of suspect transactions to follow up and a hitlist of people to interrogate question.
“Auditors can now produce these reports right down to specific individual transaction level which they couldn’t do before,” Whyte says. “It will improve the accuracy all the way from the source systems to the final report and give the auditors far more insights into how these reports were derived so they can sign off it quite comfortably.”
Crucially, Whyte still believes that firms will still need their auditors. They may not need as many as they have now, though. “Auditors still need to sign off the reports,” he says. “What it will do is reduce their time to manually go through these transactions so they can deal with more clients and look at potential issues that may exist in the actual report or the revenue cycles of firms.”
Deloitte is busy turning its algorithm into a product to use with its clients. It has already been tested with a large bank on a subset of several hundred transactions and outperformed the human team in accuracy and speed by a huge margin.
Imagine within 12 months that your audit team is travelling around the city in your Acura NSX, eating doughnuts as you pull up outside the offices of your clients’ contractors.
Just because the scene is such a classic, here are the lines from Vincent and Jules as they prepare to retrieve Marcellus’ briefcase containing “the boss’s dirty laundry”.
Jules: We should have shotguns for this kind of deal.
Vincent: How many up there?
Jules: Three or four.
Vincent: That’s countin’ our guy?
Jules: Not sure.
Vincent: So that means there could be up to five guys up there?
Jules: It’s possible.
Vincent: We should have !@#$in’ shotguns
Or you can watch a blurry excerpt here.
Image credit: Pulp Fiction