1What must be given, so that AI is used sensibly?
For AI to be used sensibly, three things must be given for every project. First, there must be a problem. This sounds banal, but without a problem definition even an AI cannot find a solution. Secondly, there must not be a simple solution to the problem using classical methods. Even if AI is a powerful tool, it is only worth using it for problems that have been difficult so far. Third, every AI project needs data. This is especially true for machine learning as a part of AI. A good database does not have to be extensive, but above all it must contain the right data. Only if the data describe the problem adequately, AI can provide a good solution. Good entry points into the topic of AI are therefore exactly where the data is located. In concrete terms, this can be machine data for typical applications such as predictive maintenance, image data for optical quality control, but also FAQs for generating chatbots. Here, AI brings a new quality to digitization and can open up new interactions, especially in knowledge management.
2Which project is best suited as the first AI application?
The best first step is always where the data is already available or can be made available very easily. My heart beats mainly for technical solutions in the field of automation and I prefer solutions where man and machine interact. An ideal entry point for AI in a company is for me the area of knowledge management. In other words, the digitalization of operational and production knowledge. Often a large part of the knowledge is stored in the heads of experienced employees and is only available to the company when the respective employee is present. By means of AI, this knowledge can be easily documented and made permanently available to all employees. By observing the human interactions, the AI can additionally deepen and constantly update this knowledge. That would be Machine Learning. Here we have focused on image recognition and processing in order to be able to detect objects reliably and provide them with additional information. This results in data records for each object, which can be managed and optimized by the AI. Everyone can imagine the added value of such a solution in the quality assurance of production processes. This is, for example, also the basic idea behind Dynamian AI, one of our current IT products with corresponding AI and AR extensions.
3What does that mean exactly?
Compared to the possibilities offered by AI, classic digital solutions work within a very narrowly defined framework that is fixed by programming. The lever in using AI lies in the database, as it is much easier to expand and adapt. I will illustrate this with the example of object recognition. A classic system is programmed to recognize a certain object, let's say a pair of pliers, in order to link, for example, instructions for use or work instructions with it. In an AI system it is sufficient to be presented with images of the object to be recognized. From this it learns to recognize it by itself and can be used much more flexible for information supply, because I don't need new programming, only new images.
4How can employees get the right information?
This is a very good question and it allows us to focus on embedding AI solutions into existing IT systems or standalone products. An AI module for object recognition alone does not create added value, I always need a complete application. Ideally, AI modules complement existing solutions and automate certain tasks. In our own product development Dynamian for work instructions, we use AI in exactly this way to simplify the cooperation between man and machine. This again shows the great advantage of AI and the use of image processing. Cooperative models are characterized by strong interactions between humans and AI. They enable employees to acquire knowledge very quickly with the help of images and supplementary texts. The provision for the employees is an important field of application, because familiarization processes and the whole area of internal training can be covered. It does not matter whether it is knowledge in production, quality assurance, sales or any other department. The cooperative approach also ensures better acceptance by all those involved. Who wouldn't be happy if the long search for assembly instructions and other documents was finally a thing of the past.
5Why is AI your passion?
During my studies, I was fascinated by technical computer science, where we built and programmed adaptive robots. Today, AI is more exciting than ever before, as technological maturity has increased enormously in many areas. Today we can not only process enormous amounts of data and create impressive solutions from it. AI is getting better and better at absorbing human knowledge, supplementing it and interacting with us. This is the basis for future innovation.