Machine Learning Makes Mercedes-Benz Dream Car a Reality


Machine Learning Makes Mercedes-Benz Dream Car a Reality
There’s a supremely easy way to find the Mercedes-Benz of your dreams on the horizon, and I watched its premiere at the SAPPHIRE NOW and ASUG Annual Conference last week.

Called the car detection app, this working prototype was a pitch perfect reflection of Daimler’s evolution of the iconic Mercedes-Benz brand for a digital-first customer experience. I caught up with Robert Kriehs, senior solution architect at Daimler, who shared the entire back story.

“Whether someone is a first-time or long-time Mercedes-Benz customer, we want to provide them with a completely digital sales experience,” said Kriehs. “We’ve already connected our customers with Mercedes me services using social media and mobile. Machine learning is the next innovation, meeting a new generation of drivers where they live, work and play with personalized, immediate services.”

Whoa, What Kind of Car was That?!
The app is deceptively simple yet powerful. When someone sees a car they love, they open the car detection app on their mobile phone and take a picture. Instantly an image recognition algorithm, based on Daimler’s image data and running on SAP Leonardo Machine Learning Foundation, displays the Mercedes-Benz cars best matching the photo, along with offers from the nearest retail dealership.

“Every class of Mercedes-Benz has numerous distinctive options, and we wanted to give people the ability to quickly identify which make and model they’ve seen,” said Kriehs. “It can be difficult to understand the differences in classes in a fleeting moment when you see one on the street. The algorithm has learned the attributes of every car model, and can help you find it for sale locally.”

Partnerships are the Bedrock of Innovation
According to Frederic Koppenberg at SAP Consulting, the car detection app is a brilliant example of the power of co-innovation, in this case between Daimler, SAP Consulting and students at the University of Applied Science.

“Strong partnerships are the foundation for great innovation,” said Koppenberg. “Daimler has the automobile industry expertise, SAP has the machine learning solutions, and the students jumped in with their ideas. The team effort produced an outcome that will benefit both drivers and Mercedes-Benz.”

Kriehs characterized the app project team as agile and self-organizing, crediting success to their shared but focused goal. “You need realistic targets with an intermediate milestone,” he said. “We reduced the scope of our original, bigger idea of using the algorithm to detect damages and predict repair costs because we realized image detection would be the ideal first project to deliver customer value.”

Unstoppable Quest for Customer Intimacy
While the near-term plan is to bring the app into production, Kriehs saw image detection as the start of innovation possibilities using machine learning and AI.

“When we know which automobile the customer is interested in, we can offer related items or offers in the context of comparable cars from other manufacturers. Customers might also be interested in collecting images of their dream Mercedes-Benz – cars they aspire to own,” he said. “These are the kinds of new business models we’re exploring.”