Accelerating automotive’s AI transformation

Publish date:

How driving AI enterprise-wide can turbo-charge organization value.

In this episode, we are focusing on the automotive sector and the role artificial intelligence (AI) is playing in accelerating the sector’s transformation: from product design and testing, to production lines and fleet management, to attracting new customers who have more choice than ever before, or customers who aren’t interested in buying a vehicle, but are scanning the market for an alternative mobility solution.

Participating in this discussion we have Bernard Marr, best-selling author, keynote speaker and futurist; Demetrio Aiello, who leads the Artificial Intelligence and Robotics Labs at Continental Tyres, and Ingo Finck, Capgemini Invent Vice President of Insights Driven Enterprise.

Also available on:

 

 

 

Transcript:

Rob Waugh:

Welcome to the Capgemini Applied Innovation podcast where we invite experts to discuss how companies are incorporating innovation into their DNA and exploring the challenges they face on their way to achieving digital mastering.

On previous episodes in this series, we’ve covered topics varying from how collaboration can improve innovation and the use of blockchain in the enterprise.

I’m your guest host, technology journalist Rob Waugh, and today we are focusing on the automotive sector and the rule artificial intelligence or AI is playing and accelerating its transformation. We’re not focusing simply on self-driving cars, but we’re looking at how AI is affecting the sector from technologies such as the use of machine learning and prototyping to the use of VR and simulations to autonomous drones used in stocktaking. It’s a technological revolution which is reshaping not just how we use cars, but how we make cars and how customers buy cars.

With me in London is Bernard Marr, bestselling author, keynote speaker and futurist. Welcome, Bernard.

Bernard:     

Thank you very much for having me.

Rob Waugh:

And from Regensburg Germany, I’m joined by Demetrio Aiello who leads the artificial intelligence and robotics lab at Continental, a company which is globally famous for supplying mobility solutions to the automotive industry. Good to be chatting with you, Demetrio.

Demetrio:   

Pleasure is on my side.

Rob Waugh:

And from Frankfurt I’m joined by Capgemini’s Vice President of Insights Driven Enterprise, Ingo Fink. Thanks Ingo for your time today.

Ingo:

Hi Rob, it’s a pleasure to join you.

Rob Waugh:

Excellent. Well, let’s get on with it, Ingo, let’s start with you as the Capgemini Research Institute has published a new study which is based on 500 interviews with automotive executives across eight countries, the study is focused on AI’s role in transforming the industry and how it’s helping to unlock value. For you, what stood out among these findings?

Ingo:

Yes, Rob, let me share some of the key findings. So number one, AI in this sector is often associated with autonomous driving. However, its applications spend a number of function areas including customer relations, R&D, supply chain manufacturing and driver experience. It’s a broader way to use it as we see it.

One of the findings is that the number of companies deploying AI at scale has only increased marginally as compared to last year’s findings, despite the potential that everybody sees behind AI in growth and cost reduction and productivity. The number of companies that apply selectively AI has not moved significantly either, so it’s just growing a a little number.

However, we see OEMs being more advanced than suppliers and dealers. So, more than 40% of OEMs implement AI at scale or selectively as compared to 25% of suppliers and only 16% of dealers. We observed a pretty big investment gap between what we call scale champions and the rest of the organizations. More than 80% of these scale champions are investing more than $200 million in AI. This drops significantly to just 20% for the rest of the organizations.

If we look at geographies, we clearly see the United States leading in terms of progress around AI, UK and Germany following, but China is clearly catching up quickly. It has nearly doubled its share of scaled AI implementations from 5% to 9%,which is the most pronounced growth rate we have observed.

And last but not least, there’s no, let’s say, clear functional focus that we observe around AI. There is no particular function standing out. Implementations around digital and mobility services seem to be slightly more advanced. However, we have seen many other very impactful implementations across other functions.

So that’s Rob, in a nutshell of what we are seeing and I think we can discuss this a little further.

Rob Waugh:

Excellent. I guess the slow adoption is slightly surprising with everything that I’ve read about AI in this sector and its potential benefits. Why is that and what should automotive companies be focusing on to deliver greater value from their AI investment?

Ingo:

Well, as we see, it’s all about scaling. So automotive organizations can effectively scale AI if they decide to focus on few but high benefit use cases and then put a really strong AI governance around these a with smart investments and also investments into expertise and up-skilling.

That means three things. One is focus on fewer projects with bigger investments. We do not see a lack of ideas or a lack of use cases. Actually, in fact, we see companies working on too many use cases, often in siloed investments. We believe that a stronger AI governance will help to allocate resources more effectively.

Second, one part of this study was what we call the up-skillng gap. So we see that the so-called scale champions have a greater focus on skills development and skills investments as a key strategy as compared to other companies.

And last but not least, adopting a streamlined process to enable scalability is important. AI is not a pure technology play. It’s a multi-discipline transformation including process, innovation, operating model consideration, and also cultural aspects you need to address. So in other words, companies should not treat AI as a niche topic any longer. It needs a sustainable investment strategy to follow through.

Rob Waugh:

Thanks Ingo for that insight. Let’s move on to Bernard. First, can I say congratulations on publishing your new book, Artificial Intelligence in Practice: How 50 Successful Companies Used Artificial Intelligence to Solve Problems. So in the automotive industry, what are some of the standout case studies that show how AI can be applied in innovative ways?

Bernard:

So I see quite a lot of interesting use cases. Many of them are still in their pilot stages. In my book I try to feature 50 different companies across all sectors from retail to media to services including manufacturing and automobile. The ones that are relevant to this conversation today are probably Baidu. They’re actually one of my AI trailblazer companies. I’ve picked 10 companies including Alibaba, Alphabet, Apple, Facebook, Tencent, and Amazon to showcase that they really have embraced the AI revolution and have really scaled up AI across the enterprise.

Baidu is the first one that is now working very hard on a self-driving car platform. What was interesting in the report was that actually you said that that there are lots of opportunities beyond self-driving cars and I completely agree.

So in my book I ought to look at BMW and how they’re using AI in different parts. One is making their products, their cars, smarter. The other one is actually making their processes smarter. They are partnering with lots of third party organizations to improve their AI capabilities from their parts tracking down to their virtual assistant Emily, in their Rolls Royce brand, which is quite cool.

Then another company that is included in my book is Volvo. Volvo has just released self-driving bus, but they’re also focusing on driving safety and they’re focusing on getting consumer feedback, so having sensor technology in their car to understand how a customers are actually using their cars,

Especially around entertainment, they did some interesting work where they’re now trying to figure out what of the entertainment features we actually use in the car because at the moment, this is all done by some level of customer research, but do we really need all the connectivity? Do we need to … What music feature? What kind of entertainment features do we actually use? So this is again, they are now applying AI and big data towards to help them track this, which is interesting.

Another company that is featured in my book is Daimler. Again, they’re experimenting with lots of different ways of bringing AI into their business. One core example is their car detection app. This basically uses machine vision. So if you found, let’s say, a Mercedes car that you really like. You take a snap of it and then the machine will tell you the exact specifications so you can order the order the same car if you want to.

But they’ve also gone beyond this. Lots of car companies are now investing in startups. They’re aqui-hiring their AI capabilities into their organization. An interesting investment that Daimler made recently is Volocopter. This is actually a self flying passenger drone. The government in Dubai for example, is trying to get this program up and running by next year. They want to have to have this autonomous flying drone service and Daimler with Volocopter, are one of the forerunners there. They actually took the crown prince of Dubai on a test ride, on a test flight recently, which was cool.

These are some of the use cases that are featured in the book.

Rob Waugh:

Excellent. So there are quite a few companies who are being adventurous with this sort of technology.

Bernard:     

Yeah, absolutely. Another car company I’ve been involved and are featured in the book is Tesla. Obviously, it’s probably the one car brand most of us would more readily associate with AI and big data.

Rob Waugh:

Thank you very much for that, Bernard. Moving on Demetrio, I’ve read that you’ve described your role at Continental as providing you with a privileged position to contribute to the future of mobility. And in terms of AI, what are the case studies that you are most excited about?

Demetrio:   

Yeah, that’s true. I’ve said that. Yeah. The reason is that for me, mobility has always been a synonym for freedom. But when you read that each year, people in Rome, I was born in Rome and that’s the reason why I’m mentioning them, spend more than 250 hours stuck in traffic jams. And I’ve read as well that you Londoners spend more than 120 hours looking for a free parking spot. You realize that for many of us, then mobility is more a synonym of pain than or freedom, yeah?

Like in many other verticals, which face an increasing complexity and reduced efficiency, automation is the name of the game. To automate mobility you need AI definitely. And for this reason, I consider myself lucky to work in this position in a high-tech mobility company like Continental in these days of profound transformation.

In our labs, we teach machines to understand the mobility world, the humans with their intentions and emotions, the environment with the objects, obstacles and free spaces and even other machines. All of this, of course, is very exciting. I would say it’s a little bit like having little kids again who every day learn from their experiences. The difference is, of course, the speed because yesterday they were still babbling and tomorrow hopefully that will already go to the university.

Rob Waugh:

So Demetrio, can you give us some example of applications in the mobility domain which are built on top of these capabilities?

Demetrio:   

Sure. Yeah. So one of the first programs that we addressed was how to reliably detect pedestrians even when they are partially hidden by parked cars, trees, or walls, on the road. Applying key point detection technique to the camera images, we found out that we were able not only to detect any person in the scene, but also to get their pose behind obstacles.

With a little adaptation, this technique has been used for other features like driver monitoring, it’s used to understand the potential level of the driver, as well as the optimal passive safety strategy in case of an accident, but also right turn assist for trucks and interpretation of traffic policemen gestures.

Another interesting capability enabled by machine learning methods is the capability to predict future events and patterns. We use techniques like LSDM and GRU to predict the expected range of an electric vehicle adapted to the driving style of the actual driver, as well as to the movement intention of traffic participants; does this car want to cut in or not? Does this pedestrian want to cross the road or not, these kind of things.

But also the demand of raw materials for our production facilities. So this is the fascinating part of the job, that the possibility to reuse the same skills in a number of diverse applications both in the area of new product development and in the area of process automation as as Bernard had pointed out.

Rob Waugh:

Well thank you very much for that insight, Demetrio. Looking at this Capgemini report, I think it turns on its head one of the things that people imagine when they hear the word AI, is that it’s going to kill off people’s job. The report in fact, found that 100% of the respondents said that AI was creating new job roles. Now this is something that I write extensively about. That’s good news because I think that in some sectors that artificial intelligence definitely does have a negative impact on the job market. So I’ve got a question for Ingo here, which is what do you believe is the one priority auto companies should be focused on in the next year when it comes to AI?

Ingo:

Yeah. So if you allow me maybe to to share not one, but maybe three ideas to share. One is we talked about AI improving technology and safety of cars a lot and improving mobility a lot. I think it’s definitely important. However, there’s still huge potential to use AI to strengthen customer relations of car companies, which is also an important strategy element. We know from other research that a big majority of people, of consumers, are actually ready to interact with AI in sales situations when they are being made aware of it. Actually more than half of them would actively seek advice from a recommend engine to buy something complex as a car. So there’s huge potential we see that should be taken forward. That’s the little bit boring side of it.

The availability and quality of data is still the key success factor for AI scaling. So it’s a matter of fact that improving the effectiveness of data management is still a must have for the entire organization, not just IT. This fact will not go away. It has to be addressed.

And last but not least, technology will make AI far more accessible to a bigger number of people than today. Also for job roles that are not very much associated to data science as of today. When we speak about up-skilling, it’s not just about developing or creating or let’s say, training new data scientists or AI experts. It’s also mainly about roles that are less data savvy today to make them more acquainted with AI technology.

Rob Waugh:

Excellent. Thank you very much for that, Ingo. Now Bernard, listening to Ingo’s perspective based on his experience working with global automotive companies, what advice would you give to companies looking to scale their AI in this sector?

Bernard:

Yeah, I completely agree with all the points Ingo made. For me, it is about identifying your strategic use cases. My recommendation is that companies identify two or three really strategic use cases. So you look at your business model and then you identify where do I have some of the biggest challenges? Where can I gain some of the biggest benefits from using AI? And this might involve looking at design and development where you use AI in your prototyping and you use digital twin technology to test some of the designs. Or it might be around sales and marketing where car companies can get the 360 degree view of how their consumers are actually using their cars or using it on augmented and virtual reality and ,the sales process as well as their manufacturing and supply chain.

Obviously, we’ve seen this whole huge investment in industry 4.0 but I see more can be done around predictive maintenance and other parts in optimizing the supply chain.

So, for me, it’s about really identifying what are your strategic priorities? So, you then look at in terms of customers, understanding customers, developing smarter products and smarter cars. Then it’s about business processes. I feel that this is where car companies and automotive companies have the biggest opportunity, looking at their own business processes, looking at how they can use AI to produce better cars, to develop better cars, to customize them, applying robotics in this process.

And so, my recommendation is identified two or three major strategic use cases, but also identify maybe two quick wins and then focus on those rather than what I see across the industry is a lot of pilots across lots of different areas that are completely unconnected. Then as Ingo said, we need to get some of the basics right. We need to focus on some of the biggest barriers to AI adoption, which include things like having the people with the right skills and the right competencies in the organization and having the right technology and data in place to do this.

Rob Waugh:

Excellent. Can you give us an example of the kind of quick wins which companies could focus in on?

Bernard:

There are so many. For me, the quick win is using AI for example, in your car design. I’ve done some work with Autodesk. They are one of the computer aided design companies. What they now do is they use AI to help their designers come up with completely novel designs that hopefully no human being would be able to come up with. This is a quick win.

Another quick win is around security. So, using AI to actually help you protect your IT system from cyber security as well. Lots of really quick wins you can implement that usually generate a huge business benefit but are not complex and they allow you to demonstrate that AI is actually making a difference.

Rob Waugh:

Great. So Demetrio, true innovation, the one thing that’s certain about it is that it’s constant and it always changes. So can you tell us about Continental’s AI and robotic strategy and specifically how have you stayed agile and ensured that you’ve stayed ahead of the market?

Demetrio:

Well Rob, as you know, our vision is to be an AI empowered company. With the word empowered, we want to stress our belief that AI will not substitute but rather extend the human capabilities. We apply these kinds of superpowers to our products to make mobility easy and fun again, and also through our processes to free all Continental employees from boring or repetitive tasks and give them more time for creativity.

To be fast in the implementation of our strategy, we need to be always at the leading edge of the research in AI. For this reason, we have set up and continue extending a collaboration network with the most renowned universities and research institutes all over the world. Moreover, we need to have a high conversion rate between innovation and digitalization. We achieve it by engaging with our internal customers very early in the innovation pipeline. We talk with them about their pain points and continuously ask for feedback to understand whether we are addressing them in the right way. Because you know, there is nothing slowing down the innovation like a solution in search of a problem rather than vice versa.

So in a nutshell, we tried to stay ahead of the market through the collaboration between the best talents in the AI research field and the best talents in the mobility industry that we hope we have already on board.

Rob Waugh:

Well, thank you very much for that. For me, Rob Waugh, that’s all we have time for today on this episode of the Capgemini Applied Innovation Podcast. It’s clear that every part of the automotive industry can benefit from AI. It’s got benefits stretching from how vehicles are designed to how parts are sourced, how vehicles are made, how they’re marketed, how they’re sold, how they’re leased, how they’re rented, and how they’re driven. It’s not just self-driving cars. AI is already pervasive across every aspect of this industry and it’s got the potential to do so much more.

I would like to thank Bernard, Demetrio and Ingo for joining us today. I’d also like to thank you for listening in from wherever you are in the world. If you enjoyed this episode, please share it via your social channels. Subscribe to the podcast series and please, please rate the show. This will help other people that are interested in innovation and AI find this episode.

If you have topics that you would like covered on future episodes, please leave a comment or get in touch. Search on Capgemini.com for the term perform AI to learn more about the companies across multiple sectors, innovating using artificial intelligence. We will also include a link within the description of this episode to Capgemini Research Institute study, which is called Accelerating Automotive’s AI Transformation: How AI at Scale Can Unlock Many Millions of industry Value.

Thank you very much, and we look forward to you joining us for a future episode of the Capgemini Applied Innovation podcast.

Goodbye for now.

 

Related Resources

AI

Collaborative intelligence for knowledge augmentation

CLICK platform enables the enhancement of system intelligence through collaboration of...

AI

Electric Vehicle Charging Infrastructure Development

Will there be enough charging points to support electric vehicle rollout in the next decades,...

AI

Emerging AI trends in Retail Operations

While most retailers are focusing their AI efforts on sales and marketing, there is a...

cookies.

By continuing to navigate on this website, you accept the use of cookies.

For more information and to change the setting of cookies on your computer, please read our Privacy Policy.

Close

Close cookie information