Deploying an Artificial Intelligence (AI) in a corporate business can be a costly endeavour. The payback period may be slightly longer. But longer term, the benefits of implementing these technologies would by far exceed the initial investment, says Prof Tankiso Moloi. He is Professor of Accountancy at the University of Johannesburg (UJ).
Profit centre owners ask themselves how sustainable they are, and how competitive they will be in three to five years. But they also need to take a close look at how AI can help them become faster, ‘see’ better, and be more agile, he adds.
Getting ready for AI
“Artificial Intelligence technology makes machines intelligent. In some applications AIs are more intelligent than human beings,” says Prof Tshilidzi Marwala.
Marwala is a globally-recognised researcher in Artificial Intelligence at the University of Johannesburg. He is the author of a series of books.
“Many AI’s are now used in industry. For example, technologies producing goods and services, or working in call centres. Some call this the ‘industrialization of decision-making'”, he says.
“Because of this, the very nature of a firm is changing. The very nature of markets are changing as well. AI technology makes markets much more efficient. Also, suppliers and customers have more similar information. For these reasons, AI’s impact on industry will keep on increasing,” he adds.
Moloi and Marwala wrote the book Artificial Intelligence and the changing nature of corporations: How technologies shape strategy and operations.
In the book Marwala and Moloi show how firms can build AI into their strategy-making and strategy implementation.
To be competitive, businesses need to prepare themselves for implementing AI, says Moloi. One example is that the interactions between business units need to be good enough before AI is even considered. Once a corporation is ready, it can gain significant benefits from AI deployment.
In the book, they discuss how to use three AI technologies that are already re-shaping large corporations. These are Machine Learning (ML), Natural Language Processing (NLP) and Robotic Process Automation (RPA).
Why AI can supercharge fraud detection
Detecting fraud is a classic example of sooner is better, and where AI can make a big difference, says Moloi.
Let’s say you’re in a corporate business and some transactions don’t pass the smell test. You may have to wait quite long for a traditional internal control system/internal audit to take place and uncover that fraud.
When the auditors arrive, they sample the financial records. Usually, there isn’t enough capacity to look at all the records, so they take a scientific sample, and base the audit on that. The thing is, the sample may or may not contain the traces of fraud to be successfully recognised as fraud.
But asking an AI to process every financial record for a long period is probably still too expensive for most businesses. Combining AI with traditional tools may be better.
“What you want to do is focus on areas where auditors say the financial controls are weaker than they should be,” says Moloi.
“Statistical modelling without AI can help point out these areas. Then you direct the machine learning AI to examine every financial record in that area for traces of fraud.”
This isn’t something human auditors can do, but a well-instructed AI can march through it all.
Some in industry report AI deployments that detect up to 90% of the incidences of fraud in a particular area, he adds.
“Recruiting the in-house specialists to ‘drive’ the AI in-between visits from the auditors is still challenging. But it is a huge advantage of knowing within hours, rather than months, about possibly fraudulent transactions.
“Also, there are major developments taking place: We’re seeing AI’s deployed by the global accounting and auditing firms, and it is expanding.”
As people say, ‘the new oil is data’. Sustaining competitive edge now requires using AI in combination with existing tools, he says.
AI improves ‘vision’ and decisions
Humans still beat AIs hands down in recognising moving objects in real life. But human brains falter when confronted with several rapidly-changing variables in high-stakes scenarios.
Yet corporate strategy and decision-making depends on exactly this.
“Companies tell us, that not using AI can mean losing competitive edge. What AI gives them is the ability to identify hidden patterns and trends in their own data, patterns they could not see before,” says Moloi.
“Without seeing these trends, they are unable to plan at the same level, or to move the organisation to where it needs to go.”
In future, AI is likely to shift much decision-making in a firm from people to systems, says Marwala.
The primary goal of AI is to do away with the errors that human behaviour tends to introduce into critical decision-making. We humans are irrational, inconsistent and can be prone to biases. We tend to take shortcuts based on our emotions, he says.
“At the moment many AI’s support humans in decision-making. But I have no doubt that gradually, more decisions in large firms will be made by systems, and fewer by people, in large firms.
“We have a classic example where this is happening right now. Autopilots fly large airplanes during most of a flight. Autopilots make many decisions much faster than human pilots can. And they use far more information than humans can process at the same time”, says Marwala.
Getting AI buy-in
“Sometimes the environment or circumstances we’re in pushes us into the next thing we have to do,” says Moloi.
The COVID-19 pandemic has forced some people, who are used to printing out documents, to slowly start making digital edits instead. And to start using online collaboration tools like Skype, Zoom and MS Teams.
He doubts a CFO will get any meaningful movement in the organisation, by giving staff the option to carry on as usual, or also start using AI.
“Rather say ‘colleagues you can attend a short programme on AI if you choose, since in three years we will be deploying these algorithms. You have to have a strategy to get buy-in,” he says.
Where to find the book
Artificial Intelligence and the changing nature of corporations is available in electronic and hardcover formats.
The book is published by Springer and available with preview HERE.