How chatbots, generative AI are transforming the car consumer experience

Like with other industries, car makers, too, are looking at generative AI as a differentiator.

In the case of Mercedes-Benz, ChatGPT will aid the brand’s MBUX voice assistant “Hey Mercedes". It will carry out more complex tasks such as pulling up the menu of an approaching restaurant or asking for real-time navigation inputs, while also recognising speech formats that are more natural and conversational.

Eric Boyd, the Corporate Vice President of Microsoft’s AI platform, has said the chatbot will allow consumers to ask follow-up questions.

Microsoft has a multi-billion dollar partnership with ChatGPT-maker OpenAI.

At present, ChatGPT integration is optional for the 900,000 Mercedes-Benz customers in the US, even though their surroundings arguably allow them to be more conversational with the car’s OS.

But what about India? Should Indian consumers rely on a tool that could be a distraction in a country with one of the highest road-accident rates in the world?

I found the opposite to be true. On a weekend drive, with an eight-year-old nephew and a four-year-old niece strapped in the backseat, I was taking on DJ duties while driving what was another German luxury SUV. Apple’s Siri, the only onboard chatbot, struggled to keep up with the extensive list of cartoon theme song requests from the children.

But the demands made on chatbots far exceed those of pre-adolescent audiophiles. In a driving environment that requires constant vigilance, any action that requires you to take your hands off the wheel can be risky.

Steering-mounted controls are now a staple, as are native voice command systems. But generative AI is expected to play a much bigger role, especially when it comes to safety functions.

Shailesh Saraph, the Global Head of ER&D at Tata Technologies, says chatbots will play a big role in the advancement of ADAS functions.

Chatbots like ChatGPT can democratise access to ADAS features, which are only offered by premium cars or the top-end variants of mass-market cars. Features such as active collision mitigation, lane assist and adaptive cruise control, can be made standard, thanks to ChatGPT.

“If you want safety you have to minimise driver distraction and the easiest way to minimise driver distraction is to operate the interface through voice commands. That’s where generative AI chatbots come in. They can have more complex interactions.” says Saraph, whose work at Tata Technologies is helping Tata Motors fine-tune its voice command function.

ChatGPTs ability to enhance human-vehicle communication means that the driver and the car manufacturer are better equipped to prevent a malfunction. The function is essentially called “predictive maintenance”, says Saraph.

Iyer agrees. He says such an advanced human-machine interface helps the company pre-empt mechanical or electrical issues.

“If the customer has given us access, we are able to do a tele-diagnosis of the car. Here AI plays a role because the car can predict the failure and send advanced alerts to the customer and to us, at the back end. So we can tell the customer to get the car to the workshop and keep the parts prepared,” he adds.

Privacy concerns

Automotive brands are hard at work to create their operating systems, free of big tech’s dominance.

“Our strategy is to come up with our own MB OS operating system," says Iyer who is also concerned about data privacy. "We want these features to exist only for those customers who opt for them. Even service reminders some people find intrusive. So the customer has a right to disable it.”

Should you opt for it, AI can also evaluate your driving. An idea most drivers will balk at but it has its benefits too.

“Insurance companies can customise the insurance brand for you. They can offer cheaper premium rates if you’re better (read: safer) driver,” says Tapan Barman, co-founder of Mihup, a voice command system that has been integrated into the user interface of Tata Motors vehicles.

Mihup is also a third-party voice command app that original equipment makers can integrate with their infotainment system. Unlike Siri, Alexa or Hey Google!, Mihup can recognise and interact in multiple Indian languages along with a mix of English and a regional language.

But the data accessible to a chatbot is immense.

“Look at the ECU in the car in India, it’s around 10 while in a developed market the average car has 30 ECUs. The amount of data that is being generated in real-time makes a difference,” says Saraph.

“People are worried about how their data is monetised, a lot of the car companies are looking to have their own OS. This concept of a software-defined vehicle is becoming very common.”

Beating Big Tech at its own game

For years now, Big Tech has held the global automotive industry in a vice-like grip, with conventional wisdom indicating that Silicon Valley is way ahead in R&D.

The perception has certainly benefited Tesla, which is considered a front-runner in the autonomous driving space. But Iyer disagrees.

“We are the only brand in the world to be certified Level 3 autonomous driving in the state of California. Even Tesla doesn’t have that. The perception is different for varied reasons but the fact is that Mercedes-Benz is the only automotive brand not running Beta versions of Level 3 autonomous cars,” he says.

The next step for all automotive brands is to come up with their operating system, one that can interact with Google and Apple but isn’t necessarily dependent on them.

Mercedes-Benz has announced that many of its new launches will be accompanied by its own operating system, Iyer says, adding the user interface will be customised a lot better, something which Apple or Google cannot offer.

“We have tied up with Google to ensure that with our operating system, the Google Map in a Mercedes-Benz system will look a lot different compared to any other car. Similarly, be it music, be it any other application,” says Iyer. “The MB OS will be our strategic differentiator when it comes to the in-car user experience.”

Their operating systems will also allow car makers to better integrate their mechanical systems with the software. Even for third-party services like Mihup, native OS is the way forward.

For Mihup, the clear differentiator and strategy is not just about building and supporting voice commands but also full automation.

“If you want to control AC temperature, it’s not possible with a third-party, non-native voice assistance function. The voice assistant needs to evolve as a multi-voice assistant in your car,” says Barman, adding user preferences matter.

And if a driver prefers Google to navigate, it’ll pass through the resident voice engine and then switch to Google, he says. A native OS allows for greater privacy and also bypasses its software for those who prefer Google’s services — offering the best of both worlds.

"It’s not about addressing one or two functions like navigation or voice command, it’s about providing an end-to-end function which requires deep workflow integration with the work ecosystem,” says Barman.

With a complex and varied demographic like in India, local OEMs have the opportunity to gain the upper hand, he says. Creating a multi-linguistic voice command system can make all the difference and collaborating with companies like Mihup, which has made considerable inroads in the vernacular voice assistant space, could be the way forward.

The future is virtual and that is where automotive firms will slug it out.

“In the future, we will see a lot of consolidation happening,” says Saraph.

Mercedes-Benz may be ahead of the game but others are not too far behind. General Motors is in talks with Microsoft for ChatGPT integration.

The possibilities are endless. Don’t wish to go through the car’s manual? Ask ChatGPT. Want to know about the maintenance history of your recently bought used car? ChatGPT has the answer.

Our cars are already smart. It’s only a matter of time before they get talking.