Artificial Intelligence (AI) is a promising technology. Who hasn’t heard about developments such as ChatGPT recently? But how widespread is the use of AI in business environments? And what can we expect going forward? Will everyone be affected by developments in AI? To find out the latest state of affairs, Finance on a Mission interviewed Phoebe Schmalz-Ziekow, product manager and AI specialist at Jedox. In our interview, we wanted to find out how she sees the developments in AI.

Is AI a hyped technology?

Yes and no. If you look at the hype around AI, you might think it has the ingredients for a compelling marketing story. In that sense, you could say that AI is hyped. On the other hand, AI is already present in so many applications that it’s fair to say it’s much more than just a hype. For instance, AI is already being used in customer support applications and to detect credit card fraud. AI solutions currently help organizations to optimize their supply chain and predict demand, sales, and cashflow. And if you look closely, the trend is to use AI more often in different areas of business, from workforce analysis to sales planning and everything in between.

Can you use AI in any situation?

AI isn’t a magic bullet. Nor is it a crystal ball that can predict the future for you. It’s good to realize that AI needs the right conditions to generate meaningful results. The data needs to be solid, and you need enough data for AI to see patterns. Still, the benefits are obvious, and many organizations that have started to use AI in their forecasting have quickly realized that it’s as accurate, if not more, than what humans can achieve. Many organizations rely on data science to help them understand the data. At the same time, we see “wizard-like” interfaces emerging that make it easier to perform the right analysis. These wizards help you to prepare and analyze data in a way that makes sense. I believe an important added value of AI is to help you understand what drivers impact your business, which will ultimately improve organizations’ business understanding.

Sounds like a no-brainer, so what’s preventing organizations from using AI more extensively?

There are a few key reasons why organizations are reluctant to adopt applications that use AI. One is a lack of understanding. What are the potential benefits, and what’s the business case for AI? Another barrier may be a lack of infrastructure and resources. AI requires good data infrastructure to get the best results. Organizations need to prepare for the use of AI, and that means first freeing up the necessary resources. Another reason is resistance to change. Organizations have entire teams involved in various business processes. How can they adjust to a new reality, where AI is taking over some of their activities? Also, having the right data infrastructure is still no guarantee that you’ll have enough high-quality data. As mentioned, AI may need large quantities of data to perform at its best. A final reason why organizations may be reluctant is their hesitance to invest in innovation when budgets are tight. This makes them reluctant to spend now, despite the potential benefits later.

How can organizations overcome these barriers?

First, I would say that it’s important for organizations to increase their understanding of AI. Look at concrete examples that are already being used by other organizations. Find out how they’re benefitting from using AI. See whether it increases accuracy and whether it would help you to increase the frequency of analysis, for example. Or how the organization can benefit from a positive business case in the long run. Involve your people in the journey without neglecting to address their fears. AI can help people to make their job more meaningful. A way to start is to look at the (business) problem you want to solve using AI. From there, you should be able to find better examples that will help you move forward. The next important question is to clearly state what you want to achieve by using AI. Once you’ve defined what success looks like, you can start collecting and cleaning data to support your AI solution. In that respect, be aware that many vendors out there are willing to help you on your journey. They can also help you to identify organizations that have already gone through this process. For example, Mitsui Chemicals, one of the companies Jedox works with, increased the accuracy of their sales forecast by up to 95% compared to the actual values for several product groups.

What important developments are on the horizon?

We are likely to see further advances in the development of AI, particularly in the field of deep learning. This technology will make AI models more powerful in the future. I also expect natural language processing to have a major impact on how AI is used in various business areas. Moreover, XAI (Explainable AI) will become more important in the coming year. These AI technologies make decisions or predictions that humans can understand. This runs contrary to the idea of the “black box”, which argues that it’s extremely difficult, and sometimes impossible, for people to explain why an AI model comes to a certain conclusion. This increased transparency will help to build confidence in the use of AI. It will also make the AI models of the future easier to use. In the coming months and years, we will see that, as with other software and applications, you can use AI with less support from data scientists. I also think that more industries and more business areas will find concrete ways of using AI.

How can we keep up with developments?

With the current pace of change, it’s sometimes hard to keep up with developments. I would suggest reading news articles and blogs by thought leaders in AI. There are plenty of examples on LinkedIn. Visit conferences on AI if there’s a specific topic that has your interest. And of course, people are always welcome to sign up for the Jedox newsletter for news on our developments in the use of AI.

If you want to read more about how AI and predictive analytics can help you in a role as trusted adviser you can download an e-book using the link below.