Last year, seeking to reclaim its position as Brazil’s biggest automaker, Volkswagen do Brasil, launched an entirely new sedan, the Virtus, designed specifically for Latin America’s largest car market.

One of the biggest changes was not to the engine, the trunk or the paint job, though. It came to the glove compartment. Or what might better be described in Brazil as the bookcase.

Brazilian law requires all passenger vehicles to come with no fewer than six manuals, including functionality, maintenance, security, even one just for the radio. As VW do Brasil, a subsidiary of the Volkswagen Group, rethought every aspect of their car, the designers began to wonder if there might be a way to bring the humble owner’s manual into the 21st century.

It may lack the flamethrower of KITT from Knight Rider, but the Virtus’ Manual Cognitivo is just as capable of saving the day when its driver is in a bind.

Built on a Watson-powered smartphone platform, the Manual Cognitivo responds to more than 8,000 spoken commands covering 150 different vehicular situations. Drivers can also snap a photo of 22 different dashboard alerts, and the app can explain exactly what the issue is. And thanks to its built-in AI, the system can learn and adapt over time.

The days of mysterious “check-engine” warnings or reaching for the glove compartment for anything but gloves or napkins are coming to an end.

“Imagine you could turn on the car and just talk to the manual,” said Andréa Crespo, a data science and AI expert at IBM Brazil in São Paulo.

Volkswagen and IBM have a longstanding partnership, whether building e-commerce platforms or developing mobility solutions. When VW do Brasil needed to achieve reliable voice recognition for the Manual Cognitivo—something capable of responding whether the Virtus was parked in the garage, speeding down the highway or stuck on a mountain road—the Watson team at IBM iX was ready.

“We look at AI as perhaps the most critical element of our innovation leadership strategy in Brazil,” said Fabio Rabelo, the head of digital and new business models at Volkswagen do Brasil.

His team chose a phone-based app, so owners can address issues and get answers even when not in their vehicle. The system can integrate with the car through Apple’s Carplay, Android Auto or MirrorLink.

Creating a seamless system goes beyond being able to understand commands like “How do I check the tire pressure?” or “How do I fade the stereo to the front since the twins are napping in back?”

The app also had to be personalized, recognizing the owner’s name and voice, the model of the car, even details like the paint job and wheel package. And the Manual Cognitivo also had to understand Brazil’s various dialects and regionalisms, as well as being adapted to Spanish, to serve VW vehicles across Latin America.

The Manual Cognitivo interface. (VW Group)

Yet the biggest challenge was pace. Because the car was so new, its features, and any discussion of them, had to be kept secret until the last possible moment. This left the Watson team four months to develop its program, including only a month for testing.

Not being able to talk to an AI built for voice recognition is quite the challenge.

“It’s important to have the feedback,” Crespo said. “We needed people talking and people’s behavior, this is what makes the artificial intelligence smarter.”

Even so, the Watson system achieved 89 percent audio accuracy within this accelerated testing phase, and 90 percent on visual prompts.

Ben Stanley, the global research lead for automotive, aerospace and defense at IBM’s Institute for Business Value, said the Manual Cognitivo represents the latent phases of the massive shift automakers will make from manufacturers of products to providers of services.

“They have to move to an experience-based platform,” Stanley said. “If the car is driving itself in 10 years, you don’t care about the performance, you care about the experience inside and outside the car.”

The Manual Cognitivo has certainly demonstrated the value of these technological investments. The VW-IBM platform helped make the Virtus one of the most popular cars in South America, where Volkswagen Group’s vehicle deliveries were up almost 12 percent in 2018, including a whopping 28 percent in Brazil (the best performance of any country for the company).

The Virtus helped drive VW sales in Brazil to the company’s best worldwide in 2018. (VW Group)

The success of the Virtus was a big part of why the Manual Cognitivo has since been rolled out into a trio of existing models on the continent: the Tiguan and T-Cross SUVs and Jetta sedan.

These AI features also helped VW do Brasil reassert its position as a cutting-edge automaker, with the Virtus winning more than 20 consumer and industry awards, the most of any model in 2018. And they also sparked interest among the millennial and Gen Z buyers VW covets as long-term customers.

While consumer interest was the driving factor for developing the Manual Cognitivo, it has provided ample additional benefits. Notably, customer calls and help requests to VW’s call centers have decreased over the past year, leading to manpower savings.

Nor is the innovating over. VW do Brasil and IBM are studying the real-world queries customers have been making for a year and a half now, which are inspiring new features both for the Manual Cognitivo as well as VW vehicles overall.

“The more cars they’re in, the better the program, and they can develop more features, connect to more security, get restaurant reservations, et cetera,” Crespo said.

And so this is but the first chapter in the Virtus’ story.

Was this article helpful?
YesNo

More from Artificial intelligence

Taming the Wild West of AI-generated search results

4 min read - Companies are racing to integrate generative AI into their search engines, hoping to revolutionize the way users access information. However, this uncharted territory comes with a significant challenge: ensuring the accuracy and reliability of AI-generated search results. As AI models grapple with "hallucinations"—producing content that fills in gaps with inaccurate information—the industry faces a critical question: How can we harness the potential of AI while minimizing the spread of misinformation? Google's new generative AI search tool recently surprised users by…

Are bigger language models always better?

4 min read - In the race to dominate AI, bigger is usually better. More data and more parameters create larger AI systems, that are not only more powerful but also more efficient and faster, and generally create fewer errors than smaller systems. The tech companies seizing the news headlines reinforce this trend. “The system that we have just deployed is, scale-wise, about as big as a whale,” said Microsoft CTO Kevin Scott about the supercomputer that powers Chat GPT-5. Scott was discussing the…

Generative AI meets application modernization

2 min read - According to a survey of more than 400 top IT executives across industries in North America, three in four respondents say they still have disparate systems using traditional technologies and tools in their organizations. Furthermore, the survey finds that most executives report being in the planning or preliminary stages of modernization. Maintaining these traditional, legacy technologies and tools, often referred to as “technical debt,” for too long can have serious consequences, such as stalled development projects, cybersecurity exposures and operational…

IBM Newsletters

Get our newsletters and topic updates that deliver the latest thought leadership and insights on emerging trends.
Subscribe now More newsletters