Artificial Intelligence at XXXL Group | Part 2

Authors: Klaus Puchner (Program Manager AI & Team Lead) and Samuel Heger (Project Manager AI)

Following our first article on the world of AI at XXXLutz Group, today we'll give you an insight into our current and future projects and provide you with a few examples.

From automated classification of customer feedback to automated image comparison – get ready to learn about some exciting projects. 

Our AI goals:

As for each strategic decision, we have thought about how to proceed when it comes to AI and how to deal with the possibilities and challenges that come with it (feel free to take a look at our first AI blog post for more details). Using AI and machine learning, we would like to provide the following benefits:

• reducing costs

• speeding up recurring processes and decisions

• increasing turnover

• improving customer experience

We will reach our goals with these measures:

• Process automation

We will achieve process automation by (partly) automating steps.

• Data-driven decisioning support

When it comes to making similar and repeating decisions, AI can assist. We will make this possible, by generating new insights and conclusions from existing data which will then serve as a basis for data-driven predictions.

• Improving customer value

AI can improve those areas that our customers actively perceive. This can be done for example by enhancing visual elements when searching for products or by enabling completely new search options. 

Some of our current projects

We will let you have a glimpse at three projects we are working on.

Natural language processing
Automated classification of customer feedback

At XXXLutz Group, our customers are given top priority. Accordingly, their feedback is vital. Our contribution here concerns our feedback forms on customer satisfaction that our customers fill out online. By means of AI, we can use the texts submitted by our customers as a basis to calculate the probabilities of which kind of feedback it is. The automated category assignment must then only be confirmed or corrected, instead of having to assign the feedback manually. The algorithm we have developed learns with each customer feedback and becomes better at assigning it. This enables us to react even quicker to our customers' needs.

Automated classification of customer feedback: AI predicts the likelihood of a piece of textual customer feedback belonging to a certain kind of category. This prediction improves with the increase of available data (customer feedback).

Computer vision
Automated creation of descriptive texts derived from image data

When looking at an image, people are able to intuitively deduce information from it and draw conclusions. This means that people are able to easily understand and process the objects and concepts (e.g. shades, atmosphere) that make up an image (people are able to summarise what they recognise in the image for example).

We are working on recreating these human abilities by means of various models and algorithms. In doing so, we want to accelerate processes that require human abilities and to minimise our colleagues' manual workload.

Text creation based on computer vision: AI automatically recognises, highlights and names pieces of furniture on images. This functionality is needed for example to automatically generate descriptive texts (i.e. what do you see on the image).

Computer vision
Image/ object similarity detection

This project also deals with recreating human abilities. When people compare two images or objects, they are able to intuitively recognise whether or not they are similar. Additionally, people can recognise nuances of deviation (to what degree they are similar). Recreating this human ability by means of AI offers new possibilities. In cases where this functionality can be used, there is a multitude of applications, for example for automatically comparing image data and for automatically evaluating the resulting conclusions to improve shopping experience.

Image data comparison based on computer vision: AI automatically compares image data from different databases and recognises similarities.

This is only a little part of what we're currently engaged in. We would like to play an active role in shaping our company's future and when it comes to AI, we have quite a lot of plans.

Become part of our team!

We develop all of our AI solutions ourselves. There is no need for us to use ready-made solutions from external suppliers. This allows for greater flexibility and security and for creating knowledge of great value within our company.

If this post has piqued your interest and if you would like to become part of our team: We are looking for an AI DevOps engineer. We're looking forward to your application!

 *German version to be found here