Introducing: AI - our co-worker behind the scenes

Let's start with the most obvious and at the same time the hardest question: What defines artificial intelligence? Well, there is no real answer to this question, because it would first be necessary to define where intelligence begins and where it ends. But that is not an easy task. An example: Is a dog intelligent because it learns tricks to get treats? Can it be, when it wouldn't even get the reward out of the bag without the human's help? Opinions may vary.  

When people talk about AI, they usually mean machine learning, which is a subcategory of AI and better describes what we think of as intelligent technology that is capable of learning. At OTTO, we also get machines to apply learned knowledge, for example in the area of image recognition. For example, after inputting an appropriate amount of data, software can distinguish whether a picture shows a dress or not. What might sound simple is in fact an extremely complex process. And it doesn't stop there: deep learning can then help to further distinguish between individual garments: What kind of dress is it? Is it short, long, elegant, casual, flashy or plain?

Why does OTTO need this? 

Anyone who has ever shopped on otto.de has likely received recommendations for similar or matching products to the ones they have bought or looked at. These suggestions arrive during the customer's visit to the site, i.e. in real time. Based on their current buying behavior, AI can use the data provided and, paired with its ability to recognize similar or suitable products visually, it can make appropriate suggestions.  

The fact that systems can analyze images independently, opens up further development possibilities and allows, among other things, for image-based search (visual search). Customers can upload images to otto.de and, based on this data, are shown similar products that can be bought through OTTO. The visual search will not replace the previous search by text entry, but it will be a valuable addition with enormous development potential in the future. 

So everything is wonderful? 

But does AI really only have advantages? Or can it also pose a threat - especially to human jobs? Let's look at BO, the chatbot from Hermes. BO is currently celebrating its first anniversary and has greatly expanded its capabilities and thus also its user numbers in this time. It also has a lot of potential for further development. However, as an additional contact channel alongside the hotline, Twitter etc., it is only intended to provide valuable support in the early stages of customer contact. BO is not intended to replace service employees. After all, emotions and the subtleties of human interaction are still uncharted territory for BO.

AI can provide valuable assistance

In general, the danger of machines replacing humans is more likely to be found in the realm of science fiction. However, one has to admit that intelligent technology is superior to our brain cells and neurons in certain areas. And it should be! After all, it was developed to make our work easier and to relieve us of tedious tasks. It is supposed to assist humans where they reach their limits: It is less sensitive to hazardous environments (e.g. mining), more accurate (e.g. automotive industry), stronger, more resilient and faster than humans.  

Not always desirable

But for all its often stunning capabilities, there are situations in which we deliberately do without AI. Coffee shops all over the world could theoretically save on staff and replace them with intelligent coffee machines, serving us hot beverages in any flavor at the touch of a button. But is that what people want? Don't you go to your coffee shop precisely because there's a guy with a beard conjuring up latte art on the foam to wish us a nice day afterwards? Admittedly, we'd like to have the barista on our OTTO campus - at least you can rent one from us. But that's why we have the smart systems. And they do a really good job in many areas at our company.