NOT TO LOSE BILLIONS

Polish companies still do not use solutions that could optimize their operations, lower costs, and improve efficiency. Eurostat data show that the digital gap is particularly acute in the area of AI. The result? Lower efficiency of the Polish economy. Yet change may be easier than it seems.

MACIEJ KOWALIK
expert in planning and using artificial intelligence in supply chains, CEO of Smartstock

The Polish economy has a problem with efficiency. On the one hand, the level of prosperity is rising. According to Eurostat, in 2024 Poland’s GDP per capita reached about 79 percent of the EU average. However, this result is achieved with relatively high resource consumption.

This is clearly visible in labor resources. According to OECD data from 2022–2023, a Polish worker logged as many as 1,830 hours per year, while the EU average was around 1,570 hours. This means that the average Pole needed up to 47.5 percent more working time to generate the same value of GDP. This does not mean that Poles work poorly. Lower efficiency has many sources: technological gaps, an unfavorable energy mix, problems with implementing robotics and automation, and the so-called digital divide.

Changes in the energy sector require years and investments worth billions. Robotics brings social challenges. Meanwhile, the digital divide should be relatively easy to close, especially in the age of artificial intelligence. But here, too, the situation is far from ideal.

According to Eurostat, as many as 43 percent of companies in the EU use ERP (Enterprise Resource Planning) systems, which are used to store and circulate data within the company. In Poland, however, this figure is only 38.8 percent of enterprises. The situation is even worse when it comes to analyzing the data collected in these systems. Eurostat reports that in 2024 an average of 13 percent of companies in the EU already used AI solutions, while in Poland it was only 5.9 percent, compared to over 27 percent in Denmark, the leader in this ranking.

AI instead of a crystal ball
This is a major loss, because these tools have huge potential to improve the efficiency of Polish companies. One of the challenges businesses face, which artificial intelligence can address, is advanced forecasting of demand for
a company’s products and services. To this end, systems based on machine learning and advanced mathematical methods are used. Running a business in today’s dynamic world of free markets and international trade is not easy. The number of factors that must be taken into account is enormous, and must be multiplied by the rapid pace of change. If we add trade frictions between China, the US, and the EU, it may seem that managers need a crystal ball or the services of a fortune teller to forecast the values of key factors for their business. In practice, companies use historical data, forecasting methods based on averages, or the personal experience of their employees. This is no longer enough.

The days when forecasts could be made intuitively, by gut feeling, or solely on the experience of in-house experts are over. There are too many factors, including external ones; global chains of interdependencies are too complex; the pace of change is too fast; and uncertainty is too high. Properly used artificial intelligence can help. In their daily operations, companies need many forecasts, covering many different product items, delivered quickly and with high accuracy. Only then can these forecasts support decisions that optimize the functioning of a given enterprise. By applying AI, this becomes feasible. In one of our projects, we helped a Polish company forecast production levels for as many as 6,000 different types of products across four continents.

Systems using machine learning can increase the accuracy of forecasts by as much as 65 percent compared to “traditional” methods based on calculating averages. In addition, they combine data from many sources and quickly adapt them to changing conditions. Reliable forecasting means the ability to optimize resource use, reduce warehousing and transport costs, and decrease waste related to expired production inputs or finished products.

Does AI make sense?
Recently, the media have been full of reports on a famous Massachusetts Institute of Technology study on artificial intelligence. According to its authors, valuations of 95 percent of listed companies involved in this technology are overestimated. This kind of situation is quite typical for new, revolutionary technologies, whose stock market valuations are driven by investor enthusiasm. However, the MIT report’s authors point out that around 5 percent of companies investing in AI generate substantial profits, and these are usually firms focused on solving one specific problem. Demand forecasting and the related ability to optimize inventory levels and associated costs are precisely an example of such a specific problem.

Better forecasts mean the possibility of reducing unnecessary inventory by as much as 20–50 percent, administrative costs by 25–40 percent, and transport costs by 5–10 percent. Better production planning also means less waste of raw materials, finished products, and energy. And this can translate into further profits. This is also confirmed by official EU data. According to the EU “From Farm to Fork” strategy, 1 EUR saved by reducing waste of agricultural products can generate up to 14 EUR in profits for the companies processing them. And there is a lot of work to be done. In Europe, processing accounts for as much as 19 percent of food waste, and trade for 5 percent. In Poland, these figures are at a similar level. This means losses worth billions. Under EU strategies, these volumes are to be cut in half by 2030. Whoever achieves this faster will not only meet legislative obligations, but will also gain an advantage. Time is of the essence. The competition is not sleeping.