Case Study: AI in the product
New bean-to-cup coffee machine
for the B2B segment
A Swabian manufacturer of bean-to-cup coffee machines wanted to significantly increase the degree of innovation with a new series for the B2B segment. Some of the machines were already rented out to coffee bars and restaurants.
The manufacturer wanted to stand out from the competitors technologically as well as regarding user experience. The concept was still at an early stage. Limiting factors were the price level as well as the usual generation cycles similar to B2C (timing).
The complexity of the task required the involvement of many departments and stakeholders, which eventually also mapped the future options. The technological and economic position of the project within the competitive landscape was determined jointly. The assessment and evaluation of possible individual technologies made it possible to derive specific overall conceptsincluding comprehensive customer structures (e.g. competencies, constraints) and processes as well as the target audience. With reference to technology, these concepts comprised new individual features as well as complete system architectures for new use cases and ways of production.
Analyses and discussions with the customer resulted in preserving the technological and high-quality overall orientation of the series, which should focus even more on the rental market and stand out on the market by various USPs. A data-driven pyramid leasing model with appropriate P/PM components and sensors seemed to be suitable for the use case and for the target customers.
For new machines, the algorithms are consistently improved by updates based on a continuous supply of new data and design approaches. This way, the systems always exhibit the best possible detection rates.
Due to the chosen extensible system architecture, the technological integration of new product features in subsequent generations proved to be easy. One extension under consideration is the integration of a cost-effective camera or a robust 3D mmWave radar including personalized Machine Learning models. This would make it possible to automatically detect cups and their sizes and to prevent operating errors or wrong positioning by giving feedback to the user.