Swarovski crystals are loved by many around the world. With countless high-end designer and retail partners, they adorn everything from evening gowns to cell phone covers. And people’s love of Swarovski has made it a prized collector item, particularly for figurines and jewelry.
But what happens if someone chips their crystal teddy bear?
At the recent SAP Leonardo Live event in Frankfurt, Werner Huber, Corporate IT Manger from Swarovski, explained the process.
“In the past, repairs have been an extremely manual process. A customer would bring their broken item into a local store. The store managers would send the item to the Swarovski repair center in the Austrian Alps, which typically receives over 100 things each day. Then the repair technician would have to manually identify the correct material number for the piece, find out when it was produced, what business unit it came from, and determine the exact parts required to fix the item.”
Werner continued, “Technicians up to this point have developed home-grown solutions: they manually search through Swarovski catalogues or just Google it.”
That’s why Swavorski teamed up with SAP. Together they developed a machine learning application to help automate the process of identifying and classifying broken items.
As a result, Swarovski has a new process that is faster, less error prone, scalable and able to respond to demand more flexibly. Now technicians can run service tickets, with an image of the broken product, through the machine learning application. The product is automatically categorized, and repair staff can get precise information about the product: what it will take to fix it, how much it will cost — and if it’s not fixable, the app will provide suggested substitutions that might interest the customer.
Sebastian Wieczorek is a director at the SAP Innovation Center Network, which spearheads the company’s machine learning strategy and products. He explained, “If you take the image of the bear, the algorithm understands the semantic concept of the image and returns a handful of results – all are animals, matching one or more of the concepts in the image that the algorithm was trained on.”
The application uses machine learning, but it also relies on SAP’s visual recognition technology – a subset of machine learning that mimics human vision. Applications using computer vision can automatically “see” and understand pictures and videos faster than humans could ever hope to. The algorithm can be trained on particular images: in this case, thousands of Swarovski product images in catalogues.
SAP has incorporated the functionality to recognize and classify images in an API, which sits on SAP Cloud Platform, making it easier for companies to custom-build applications like the Swarovski app.
Swarovski was founded in 1895. But like many traditional businesses, the company is finding that it needs to embrace newer technologies to keep customers happy and engaged. This is one example where digitizing operations enables the company to serve customers more efficiently. Swarovski is also exploring how technologies like artificial intelligence can help the company better understand and keep their loyal customer base.
Like the next season of Swarovski collectibles, we are eagerly awaiting more examples of Swarovski’s evolving digital strategy.