We use cookies to personalise content and ads, to provide social media features and to analyse our traffic. We also share information about your use of our site with our social media, advertising and analytics partners. Read More


Log in

How will Artificial Intelligence affect the role of the modern-day accountant? – Andrew Cachia

10 Feb 2020 10:58 | Deleted user
The Accountant – What's Next? – Winter 2020 (MIA Publication)
Artificial Intelligence (AI) has been gaining rapid popularity over the past few years, and with good reason. Although AI has been around for a long time, with some of today’s commonly used algorithms dating back to the 70s, recent advancements in computational power as well as the abundance of data available have helped AI to flourish and demonstrate its capabilities.
In order to understand how AI can provide value to accountants, we must first understand what AI is and, more importantly, what it is not. Automation and Artificial Intelligence are terms often used interchangeably and, although they go hand in hand, the concepts are distinct.
Automation is the process of replacing manual tasks with systems that can perform those tasks autonomously with minimal human intervention. It streamlines a series of processes ensuring they execute in the correct order, and performs calculations, checks, as well as the generation of necessary notifications or reports. No fancy algorithms are needed. There is no limit really to what can be automated, but it usually boils down to resources or cost. But the long-term cost benefits of automation can be huge, reducing the grunt work performed by employees, allowing them to focus on the bigger picture rather than on inputting formulas into a spreadsheet or punching in numbers on a calculator.
In fact, it still boggles my mind how going into 2020 so many companies still operate in an old-fashioned way, calculating payrolls and taxes, manually crunching up numbers on spreadsheets and so on. These are processes that an automated system can handle very well, in a much shorter time frame, and with minimal chances of human error.
Where then, does that leave Artificial Intelligence? Artificial Intelligence is the process of solving non-trivial problems using methods and algorithms that mimic the way in which human beings understand and learn problem-solving. To use AI, you must have a problem to solve, one which cannot be solved using traditional techniques as they would be ineffective or inefficient.
An example of something Artificial Intelligence is very good at is pattern recognition, especially when dealing with complex sequences spanning huge ranges of data that involve a multitude of variables. Manually sifting through all the data would be extremely impractical and traditional time series analysis usually has trouble determining the relationships between all the different variables. Even when this works, since many of these factors change over time, models quickly become outdated. AI, on the other hand, is not only capable of determining underlying hidden patterns and relationships between different variables, but can continuously monitor and update itself, learning on the go and remaining relevant to the most recent data.
This is useful for generating predictions, allowing companies to forecast financials, based not only on the company’s past and recent performance, but a whole range of other factors too, including the economic climate, industry trends, seasonality and competitor performance. Seasoned analysts can also dive through all this data and generate estimates, but AI is not only quicker and is able to pinpoint underlying correlations between the company financials and internal or external factors that analysts might not see.
AI also excels at dealing with quantifying very subjective values, with risk assessment being a good example. The days of relying on the Sharpe ratio to quantify risk are soon gone. AI algorithms analyze many factors to generate probabilities that  help = the company’s decisions.
AI is also good at conglomerating non-standardized, incoherent data from multiple sources, retrieving the desired information, and storing it in a central location. Nowadays, data such as requests, receipts, invoices or quotes come in from multiple different streams, including emails, direct messages, handwritten notes, printouts, or specialized software systems.  Why should the accountant have to go through the cumbersome process of typing details from invoices and receipts into the accounting software system, when AI can scan them and pick the correct details accordingly? Although Optical Character Recognition (OCR) systems have been around for a while, they usually require a strict, standardized format. Try feeding receipts handwritten by multiple different people to an OCR! AI can identify the required details more easily and will only get better with time.
As the adoption of automation and AI becomes more widespread, the role of the accountant will change. Instead of focusing on number crunching, accountants will focus more on value adding activities and reduce the repetitive, monotonous work . Their role will begin to shift to that of consultants and auditors. Although AI is powerful, it is not foolproof, and the output and results produced will need to be cross-examined by accountants who are knowledgeable in the domain, especially during the early days of adoption. The AI algorithms also need accountants to give them problems to solve and guide them as company objectives and policies change over time.

Andrew is an experienced software developer within the FinTech industry and is currently completing a master’s degree in Artificial Intelligence.

Suite 4, Level 1, Tower Business Centre, Tower Street, Swatar, BKR 4013, Malta 

E-mail: info@miamalta.org

Tel. +356 2258 1900