The VDMA industry association surveyed a total of 70 companies on the topic of artificial intelligence. Currently, there is a medium to high relevance for machine learning-based solutions to support business processes, particularly in the application areas of customer service (64%), design and development (54%) and production (50%). In terms of importance for products, machines (69%) and predictive maintenance (69%) and condition monitoring (65%) are of medium to high importance from a company perspective.

Around 46% of participants already use a corresponding solution for company processes or in products or services. The current application focus is mainly on design and development (14%), customer service (13%), production (13%), accounting and controlling (10%), condition monitoring (13%) and remote service (13%). However, the companies surveyed are planning to significantly increase their use in processes and products over the next three years. By 2022, for example, more than half of the companies want to use machine learning-based solutions in customer service.

ML supports process automation and communication

The initial experiences of the survey participants with the use of ML-based solutions show very different effects in internal company processes and at the customer. Particularly in non-manufacturing processes, the use of ML has already led to a reduction in personnel costs (time, costs, etc.) in around half (41%) of the companies. In addition, the four most common effects also included an increase in the degree of automation in the processes (37%), improvement in service support (33%) and communication with internal or external process participants (33%). The most frequently cited effects in the company's manufacturing processes are also predominantly similar. In addition to a reduction in personnel costs (44%) and an increase in the degree of automation (41%), the use of ML-based solutions in products or services by mechanical engineering customers primarily created the opportunity to offer new products or services (48%).

Lack of suitable personnel and data

Of the five most common reasons why machine manufacturers are not yet using ML-based solutions in their company processes or in their products or services, a lack of human resources (64%) and a lack of qualified data material for training (64%) take the top spots. However, an unclear benefit or return on investment and unresolved legal issues are also decisive factors for 43% of participants for not yet using it.

It is undisputed that projects will not work in the future without experts in machine learning or artificial intelligence in the company, as in many cases data must first be analyzed and models trained for the subsequent task. 42% of participants are therefore planning to make corresponding hires by 2022. Small and large companies alike are looking for this expansion of expertise.