Can machine learning be the finance director's new best friend?
Finance directors could benefit from an area which continues to gain traction - most notably as a security tool for emails and expenses. Sooraj Shah investigates.
The hysteria surrounding machine learning could lead to some finance directors switching off any time it’s mentioned, but there could be some very specific-use cases that FDs could benefit from by using the technology.
Machine learning is the ability for a machine to constantly learn without specifically being programmed to do so. The software uses large volumes of data to understand patterns and adjust accordingly. For example, eCommerce giant Amazon uses machine learning to understand which products people are buying, and recommend them products that other people bought.
With the number of use cases and interest growing, there are numerous tech start-ups dedicated to the field of machine learning. One of which is CheckRecipient, a London-based company founded by three engineers who graduated from Imperial College London.
The concept behind the company is a machine learning algorithm that automatically checks whether an email is being sent with confidential, private or sensitive data to the wrong recipient.
And it has gained interest: CheckReceipt has just raised $2.7m in a funding round led jointly by Accel and LocalGlobal, and already counts hedge fund Man Power and law firm Travers Smith as its customers.
Tim Sadler, co-founder and CEO of the company, told AccountingWEB, that the company was started because of the amount of highly sensitive financial information he and his co-founders saw being shared with the incorrect recipients in the M&A sector.
He said that the same kind of critical information is being held by FDs and accountants.
“Accountants will have non-public, non-published financial data and it is one of the most sensitive things that can be inadvertently leaked, and this could breach FCA regulations and could affect the share prices of a public company,” Sadler explained, adding that the finance function has obligations to keep information secure, while accounting firms would be under confidentiality agreements with clients.
Finance directors are frequently targeted by cyber criminals because they have an influence in regards to moving money from different accounts”
According to Sadler the most common digital security issue reported to the Information Commissioner’s Office (ICO) is misaddressed e-mails, and machine learning can help finance directors by automatically checking for anomalies.
CheckRecipient is also working on an anti-spear phishing add-on. Sadler suggested that FDs were frequently targeted by cyber criminals because they have an influence in regards to moving money from different accounts. Again, machine learning would be used to detect phishing emails and make this clear to the individual reading them.
But email security is just one area in which machine learning could help those in the finance function – expense fraud is another.
“While a leaked email might make the headlines, expense fraud can really hurt a company behind the scenes,” Chris Baker, managing director of UK enterprise at expense management company Concur, claimed.
Baker suggested that expense fraud was a “consistent, deep-rooted and often hidden issue”, particularly because in large multinationals finance teams are tasked with going through hundreds, if not thousands of tedious claims.
This can lead to employees being hit by both boredom and time constraints, and ultimately lead to errors.
“That’s why machine learning is so important in this area because it can spot and flag anomalies that all too often would simply slip under the radar,” said Baker.
As machine learning gets more advanced over the coming years, there are likely to be even more use cases for FDs and accountants, which aren’t limited to securing existing working methods. Instead, machine learning could be used to make better business decisions, faster.
For example, they could work closely together with the risk function to identify risks of working with a third party and they could have a better understanding of the financial data at their disposal, which could help other departments as well as the CEO from a strategic perspective. They could also help to automate manual tasks such as the series of checks necessary within an auditing process.
However, while the power of machine learning is likely to have a profound impact on the way businesses operate, as Omar Mohammed, operations and financial markets analyst at Imperial FX states, the technology still require human skills and knowledge to make it work to its full potential.
It’s up to the FD to decipher between machine learning tools that provide a solution, and those that are merely an unnecessary add-on.
Do you use machine learning in your role? Have you found it useful?