How BMW Group has embraced AI for positive use cases and to improve sustainability | AWS again: Invent

Artificial intelligence (AI) is helping many different industries and is having a particularly big impact in the automotive industry. One of the most exciting use cases is for fully autonomous vehicles, but that’s not the only area where AI has an impact. For example, Microsoft and Mercedes-Benz are working together to improve the production efficiency of cars.

At the AWS re:Invent cloud conference this week, BMW Group outlined the impact AI has had on the organization and detailed emerging use cases where AI will deliver future positive business outcomes.

In a session, Marco Görgmaier, GM, data transformation and artificial intelligence, BMW Group, said his team had built a library of thousands of data assets across the company that can be reused for analytics and AI. He said his team has been able to deliver more than 800 use cases since 2019 that have generated more than $1 billion in US dollar value. The use cases include research and development, logistics, sales, quality and supplier network.

“Our team’s vision and mission is to drive and scale business value creation through the use of AI across our value chain,” said Görgmaier.

BMW on its way to a sustainable future with some help from AI

An emerging area where BMW is now investing resources is helping improve sustainability.

Görgmaier noted that 60% of the world’s population lives in cities and urban areas, which also generate 70% of greenhouse gas emissions. What BMW is trying to do now is help city planners solve problems to help reduce emissions.

BMW is already helping with machine learning models that can predict how traffic regulations could potentially help reduce both traffic and gasoline emissions. ML models are also used to identify where there is not yet sufficient charging infrastructure for electric vehicles. Görgmaier said a lack of charging infrastructure prevents people from switching to an electric vehicle, which in turn has implications for sustainability.

There is also a BMW ML effort to help predict the impact of parking availability and pricing on driving patterns. Those patterns include commuting routes and traffic, which will also have an impact on emissions.

Generate geospatial information with Amazon SageMaker

Görgmaier said many of the urban sustainability problems BMW is trying to solve could benefit from geospatial information. That’s where BMW starts leveraging new geospatial capabilities in the Amazon SageMaker ML tool suite that was just made public this week.

One area where BMW wants to take advantage of geospatial ML is to help predict when an organization with a fleet will be able to transition to electric vehicles.

“We set the goal of training machine learning models to learn correlations between engine type and driving profiles,” he said. “The rationale behind that was that if such a correlation existed, the model could learn to predict the affinity of certain drivers for an electric vehicle based on their profiles.”

Since BMW was working with fully anonymized fleet-level data, it had to use GPS traces and geospatial data to make the correlations.

“At the end of the training, the model was able to predict with an accuracy of more than 80% how likely it was that specific fleets would switch to electric vehicles,” says Görgmaier.