Artificial intelligence (AI) technologies can open up great opportunities for companies in terms of new technical applications, digital business models and practical facilitation of everyday life, and have disruptive potential because the scientific and economic application possibilities are far-reaching. This results in multifaceted connecting points along the value chain and many even expect the number of AI applications to grow exponentially in the coming years. Experts believe that artificial intelligence will have a significant positive impact on global economic growth.
The basis for this rapid growth is the amount of available and linkable data, increasing networking and ever better computing power, which allow a greater degree of automation and individualization of products and services. This gives hope that there is also high potential here for addressing societal challenges such as climate, mobility and health. A survey of executives conducted in Germany, for example, revealed that a majority of companies consider AI to be important for competitiveness and the ongoing existence of the company. Nevertheless: AI is often perceived as so far-reaching that its implementation represents a major barrier.
Experts currently see the greatest advantage in the possibility of providing further support for automation with AI tools. This makes them the ideal enhancement for ERP systems, whose core task is, after all, to increase the effectiveness and efficiency of the company. A great opportunity in this area in particular is that it does not always have to be directly revolutionary technologies that companies integrate into their business. Instead, a step-by-step approach is recommended: Initial experience is gathered with so-called use cases - for example, how well the integration of AI into the existing company system works - thus creating the first success cases. In this way, certain processes or sub-processes are first enriched with AI technologies and employees are supported in their work - at best in order to free them up for more profitable tasks and to use their know-how in a more targeted manner.
Finally, high optimization potentials with regard to the use of resources are often the decisive reason for companies to use AI. These relate to the time capacities of employees, for example. Another example is material planning, which can be further optimized with AI in a more dynamic, agile manner and across locations based on seasonal aspects or fluctuations in demand. Other processes along the value chain such as sequence calculations of orders, optimization of delivery dates and material deployment, or logistics planning are also thinkable, because another benefit of AI is that it supports decision-making. Interrelations between data are automatically tapped and thus a solid basis for decision-making is created, which leads to ERP systems further becoming an intelligent integrated business solution.
AI thus certainly has the potential to expand human capabilities and thus help us gain new insights. However, there are still many uncertainties alongside these possibilities:
A central issue is the concern about the extent to which ethical principles are upheld. Depending on the database used, there is a possibility that AI applications will also make discriminatory decisions. Namely, if the training data was not processed appropriately during machine learning. In this case, a recommendation for action can deviate completely from the understanding of values - this is called bias. As an example, in an application process, only males between 25 and 35 could be considered because the highest efficiency is derived from the historical data.
But also the manipulability - the so-called robustness of models based on AI technologies - is a central factor that repeatedly raises fear and mistrust.
Not least, the topic of the General Data Protection Regulation naturally also plays a significant role, since machine learning is giving rise to new methods for linking data and AI applications are thus bringing together data that was previously not linked. This so-called "data linkage" increases the risk that individuals or companies can be identified.
Accordingly, an important point on the implementation agenda is to answer these uncertainties. If it is made transparent how the requirements for AI systems, namely fairness, interpretability, robustness and security, governance and ethical issues are implemented, a first major barrier with regard to acceptance for the use of AI-driven solutions in the company has already been overcome. This is because fair behavior of the AI application towards all stakeholders, for example adaptation to the needs of the users, an understandable, reliable and secure mode of operation, as well as the protection of sensitive data, strengthen trust in the AI application.
And since spring 2020, the German Federal Ministry of Labor and Social Affairs with its "AI Observatory" project has also been focusing on how "responsible use of AI in the world of work and society can succeed in accordance with the goals of the German government's AI strategy." This is intended to provide assistance and empower and strengthen players in dealing with artificial intelligence.
This makes the task clearer for companies when it comes to the next steps in operational digital transformation. It is not a matter of implementing digital innovations in the company for innovation's sake. Nor is it an item to be checked off the next agenda. Rather, it is about using technical progress sensibly for the further development of the company - there is almost infinite potential. The decisive task is to filter out the relevant options in accordance with the company's vision and mission.
And since the integration of these technologies into corporate processes is just beginning, it is important to keep an eye out for suitable applications and to open up to the possibilities and opportunities.