AI in business: Taking a longer term perspective

Paul Geertsema joins Vlerick as Professor of Accounting and AI

“Artificial intelligence and machine learning have left the lab”, says Paul Geertsema. “Now it’s time to do some serious thinking about the real-world application of these technologies. What are the challenges? What could go wrong? How can we mitigate the risks? There’s a lot of work to be done.” Recently appointed professor of Accounting and AI, Paul is ready to roll up his sleeves and tackle these questions head-on.

web-insights-paul-geertsema
  • Associate professor of Accounting and AI
  • Formerly Senior Lecturer in Finance at Auckland Business School
  • Applies advanced AI approaches and methods to solve problems in accounting and finance
  • Naturally curious, loves to read and take walks to enjoy nature

Paul joins us from Auckland Business School where he was a Senior Lecturer in Finance. Before moving into academia, he worked in the financial services industry. Today, he combines his expertise in IT, accounting and finance to explore how advanced AI approaches and methods can add value to business.

Enter generative AI

A hot topic in his area of expertise is undoubtedly the emergence of large language models1 such as ChatGPT, and more generally, of generative AI, which has fundamentally changed the landscape of machine learning. “Two years ago, you could train models to do a specific thing, such as playing chess, translating from English to German or image recognition, which they ended up doing reasonably well, sometimes very well”, explains Paul. “What has changed with generative AI is that these large language models can do many different things and do them effectively. For example, the same model can solve mathematical problems, translate and explain how to bake a cake or plan a holiday. More importantly, these models can perform tasks for which they haven’t been explicitly trained.”

The computer nailed it

Now that we know what AI and machine learning tools can do, the challenge is to apply them in real business contexts. This is exactly what Paul and co-author Helen Lu, who also happen to be husband and wife, have been doing in their research. One recent example is their use of machine learning to value listed companies. Paul explains: “Traditionally, consultancies and investment banks value a company by opening up a spreadsheet, plugging in all the data, doing a bunch of calculations and out comes a number. That number, which depends on the assumptions they use, is then compared to the values of peer companies. We proposed a different approach: why not take all this input data and train a machine using AI to figure out the relative valuation, using the huge number of examples available? For every listed company, we have a lot of information such as earnings, cash flow and book value, as well as the stock market valuation. We used this data to train the machine. And what we found is that machines can actually do surprisingly good valuations. What was even more remarkable was that the machine learned to value companies in a similar way to human analysts. Basically, the computer figured out the rules for valuing companies on its own.”

Their model, based on historical data, provides an estimate of a company’s value that can be compared to the company’s actual stock price. If the estimate is lower or higher, it suggests that the company may be over- or undervalued in the current market. “We noticed that companies marked as overvalued had below-average returns in the following month, with their prices falling. On the other hand, those flagged as undervalued saw their prices rise, resulting in higher-than-expected returns. This suggests that our model’s valuations behave like fundamental or intrinsic valuations.”

Not just for the few

One of Paul’s passions is sustainability: “It’s about global climate change, and how we can grow our economy while taking into account the environmental impact. And it’s about creating organisations that benefit everyone, not just a few.” He has equally strong views on AI and machine learning technologies: “They shouldn’t be exclusive to a few or only understood by tech experts. I want to bring these amazing technologies into the real world where they can benefit everyone. Businesses should use these tools effectively to increase efficiency and create more value for all their stakeholders.” And he emphasises: “We need to consider the social justice implications of these technologies and ensure that the prosperity they create is fairly shared.”

Walk the talk

With its clear strategic focus on sustainability and digital transformation, the School resonated with Paul’s values and interests. “What attracted me to Vlerick is that they’re not just paying lip service. Sustainability and digital transformation are actually embedded in everything they do, from courses to operations”, he says enthusiastically, adding: “I also appreciated its agility as a smaller organisation. There’s less bureaucracy, which makes it more entrepreneurial. If you have a good idea, you don’t have to go through a years-long committee process to get it done.”

Cross-fertilisation

At Vlerick, Paul will combine his research with teaching in the Masters, MBA and Executive programmes. He finds this combination ideal. “If you only focus on research, it can quickly become a rather sterile exercise. Students will challenge you, especially postgraduate students who already have some work experience. This way, as a researcher and teacher, you’re also exposed to new ideas and teaching becomes more of a two-way street.” He pauses. “That said, research is very important to me and I’m looking forward to working with my colleagues at Vlerick to see how we can apply these technologies.”

Curiosity in the driver’s seat

“Finding out how things work is its own reward”, according to Paul. In fact, it was this curiosity that drove him from finance to academia. “While finance is a challenging and rewarding career, one of the attractions of academia is that if you find interesting questions, you have the opportunity to explore them in a way that is not possible in commercial organisations. I kept seeing all these fascinating problems, but I had to focus on the P&L and making sure the trading desk had a good year. So that was my motivation to change.”

Asked what he wants to achieve, he replies: “We’re currently in the midst of one of those great industrial-revolution-type transformations. Over the next 5-10 years, the impact of AI and machine learning will be significant, similar to what we’ve seen with electricity or the internet. I want to be part of that transformation and help shape it.” And he points out: “Not just on a day-to-day basis, but strategically - thinking about where we want to be in 5-10 years, what kind of world I want my children to live in when these AI tools become even more powerful than they are today.”

Profile

  • Associate Professor at Vlerick Business School
  • Full member of Chartered Accountants Australia and New Zealand
  • Founding Board Member of the AI Researchers Association (2021-2024)
  • Senior Lecturer in Finance at Auckland Business School, New Zealand (2014-2024)
  • PhD in Finance at the University of Auckland, New Zealand (2011-2013)
  • Founder and director of North Shore Consulting Ltd (2009-2024)
  • Associate Director at Barclays Capital in London, UK (2003-2006) and Hong Kong (2006-2008)
  • Master of Business Administration at London Business School, UK (2001-2003)
  • Bachelor of Accounting at Stellenbosch University, South Africa (1991-1994) and Bachelor of Computer Science at the University of Auckland, New Zealand (1995)

1 Large language model: a type of machine learning model that has been trained to understand and generate human-like text. Large language models fall into the category of generative AI. Generative AI is not limited to text, but also includes models that generate images, music, speech, or other data.

Get in touch!

Paul Geertsema

Paul Geertsema

Associate Professor of Accounting and AI