Professor of Management Accounting and AI-driven Finance
The internet, the smartphone and the cloud have already been game-changers – but the technology that will really turn our lives upside down in the coming years is undoubtedly AI. How can artificial intelligence also make a company’s financial department more efficient? And what roadmap should organisations draw up for this? Kristof Stouthuysen, Full Professor of Managerial Accounting & Digital Finance at Vlerick Business School, shed light on these matters during the CFO Day.
A participant wondered: “Is AI technology developed enough to really help organisations?” Professor Stouthuysen was crystal clear: in his opinion, anyone who still has doubts about AI is already too late. Even more: “Today's CFO is confronted with technology that makes better decisions than he or she does. Research by Stanford University, the so-called AI index, shows this clearly. Whether it concerns generating text, drawing up plans, solving mathematical problems, and so on – AI often does better than humans. Humans do have a slight advantage when it comes to very complex environments in which context plays a major role. But it’s abundantly clear that, even there, it’s only a matter of years before AI takes the lead.”
In the coming years, AI will become exponentially better, more accurate and more powerful, the professor said. The reason for this is obvious: the technological components that constitute the power of AI are getting better and better. “The volume of data that we can store today has increased incredibly. The big challenge in the coming years will mainly be in unstructured data, such as text and audio. This evolution goes hand-in-hand with the evolution of algorithms. For example, it has only been since 2011 that we have had algorithms that are powerful enough to recognize the difference between a dog and a donut. Add to that Moore’s Law – which states that the power of microprocessors doubles every 2.5 years – and you have a perfect cocktail for pushing AI to the next level.”
From $3.2 to 200 billion
So, it’s not surprising that the business world is fully embracing AI. According to Goldman Sachs, investments in AI amounted to around $3.2 billion worldwide in 2013. In 2020, that was 48 billion dollars; and in 2024, that will amount to $130-150 billion. The current predictions for 2025 are around $200 billion. “Science recently published an article about an experiment that showed that people simply perform better when they have AI at their disposal than when they do not,” says Kristof. “They had professionals from various domains (marketing, HR, finance, etc.) carry out a typical assignment. The guinea pigs were divided into two groups: those who were allowed to use ChatGPT, and those who did not have ChatGPT. The first group performed better in all areas. AI can clearly help improve human performance.”
7 steps
Professor Stouthuysen notes that finance is a somewhat separate world with its own DNA. So, simply investing in AI and throwing money at it is not enough. According to him, smart companies should follow a 7-step plan.
“The first step is to draw up a digital strategy. Technology should be a means, not an end. Try to see yourself in three years: who are your stakeholders? What do they want? And how do you want to raise your productivity? By doing real-time reporting, for example? Or by issuing more accurate forecasts? Or by relying less on manual input? Link that to the technology.”
The second step is to apply the data. “Finance has always been quite good with data, but it’s important to use it creatively and bring structure to it,” says the professor. “A challenge is the amount of data – which will only increase now that operational and external data are also becoming important. A second challenge is data liquidity: the flow of data across different departments. And finally, there is the challenge of data governance: that’s something that the CFO must be concerned with as well.”
Wanted: Skills
The third step in the process is to employ new skills. Kristof: “IFRS, VAT, valuation rules, … We all know that. But what does new technology entail? We don’t know so much about that. Therefore, education and training are a must.
The fourth step is to ensure that your AI remains explainable. If AI makes a suggestion, humans must still be able to determine why that suggestion is being put forward. Your customers and colleagues will expect that. If a bank customer does not get a loan, the bank must be able to explain how it came to that conclusion. That goes for technology too; it’s important not to put all your eggs in the basket of the big technology players, but to keep an open mind.”
The fifth step ties in with Step 3. “Companies need to install a data-driven culture. Switching to a technology that makes suggestions is not easy. The CFO needs to understand the technology him- or herself. That is certainly the least the employee can expect. He or she doesn’t have to write program code all day, but basic knowledge is absolutely necessary.”
The sixth step is to accept that there will be pitfalls and that not everything will go smoothly, the professor warns. “You have to learn through trial and error. That learning curve is simply a fact, the know-how doesn’t happen automatically.
Finally, the seventh and last step is to ensure digital inclusivity. Generative AI, for example, is very often trained on Anglo-Saxon data – which in fact are not always our data. A CFO must also take the single source of truth into account. Give the same data to ChatGPT, to Gemini from Google, and to Llama from Meta, ask all three of them to make a profit-and-loss account, and you will probably get three different reports. Finance must be able to intervene in this, from a human background.”
Professor Stouthuysen says that anyone who wants to start with AI does not have to immediately do it on a large scale. Small steps work fine. “You don ’t have to go from zero to a hundred right away. AI linked to smaller data volumes works as well. Moreover: the technology is cheaper than ever. There is a lot of open-source software available. AI Analytics whereby you use your own data and your own models, certainly also has its value. You don’t have to jump on the Generative AI train right away. After all, you are then also dependent on what the big tech giants can do with it, and that’s not always an advantage.”
Still steps to take
One of the finance professionals who has already taken Kristof’s advice to heart is Veronique De Roose of stock exchange listed food producer What’s Cooking? (formerly Ter Beke). “As far as AI and technology are concerned, we have scored fairly average up to now,” she says. “We knew we still had steps to take. IT and Finance have each drawn up a separate roadmap for this, albeit in close collaboration.”
Ultimately, that effort had to lead to several things, says Veronique. “We wanted to make our current infrastructure future-proof, handle data governance better, and make IT a more integrated department. IT is no longer an isolated island but a partner for the business, pulling in the same direction.”
What’s Cooking? examined all its processes and asked the question for each one: how can we improve this? “We created a map of all of our operations and scored the processes: are they lean and mean? Do we have the necessary skills in-house? From there, IT-business projects were developed, each with its own priorities.”
Ambitious roadmap
The roadmap that What’s Cooking? used was quite ambitious, in terms of both time-frame and content. “We are currently busy selecting the tools and designing the BI system,” says Veronique. “We are also starting to analyse the data and prepare the data for internal and external reporting.”
We created a map of all of our operations and scored the processes: are they lean and mean? Do we have the necessary skills in-house? From there, IT-business projects were developed.
By the end of the third quarter of this year, the CPM tool should be configured and go live, with the necessary training being factored in. Towards the end of this year, the business units’ reports will be analysed: what are the priority KPIs? And should KPIs be added? This should pave the way for integrated reporting and, later, budgeting and cost price reporting. By then, we will be mid-2025.
Veronique admits that this is an ambitious plan. “We are currently in the process of divesting one of our divisions; and there is also a major IT plan on the table to switch to cloud infrastructure. So, for the moment, we have pressed the pause button. But it is clear: this plan is fully supported by the board, and it’s important to push the accelerator as quickly as possible.”
Author: Frederic Petitjean.
This event was organised in collaboration with House of Executives.
Eager to know more?
Discover our three-day Digital Finance programme – and get the skills and technologies to lead your team to the future.
Project Manager New Business