AI/Automation
What More Can Automation Bring To The Automotive Business?
We’ve all heard Mary Barra, CEO, General Motors, repeatedly refer to how the automobile industry is changing more today than it has in five whole decades, as it awaits a revolution in personal transportation. The change this is bringing is not only fundamental but deep-rooted, caused by actors on the outside such as Tesla and Uber, slowly but surely making both driver and owner redundant with their autonomous vehicles and ride-sharing model. The impact is starting to show.
For the seventh month in a row, U.S. car sales in July this year continued to decline. And predictions are that carmakers are likely to sell about 17 million vehicles or fewer this year compared with nearly 17.6 million last year. This has serious consequences for the industry, which although highly efficient in its manufacturing methods (read just in time, LEAN etc.), is dated in its selling practices: basically, instead of producing against an order, car manufacturers churn out cars based on a demand forecast made months in advance and then flood the dealerships. It doesn’t take long for the inventory to pile up and because a new car can only remain new for a year, unsold stocks are flogged as old, at a discount or easier credit. Amidst all this, automobile manufacturers are investing enormous sums in autonomous vehicle technology, with no sign of payback. The only way they can sustain that is through cutbacks and layoffs. GM has already laid off workers, and there’s been talk all year that Ford will follow suit.
As traditional automakers struggle to stay profitable in the short term, and alive in the long, we believe that automation - in each of the three most important entities in the ecosystem, namely manufacturers, financiers and dealers - could provide the answers.
Physical automation of tasks and processes
The automobile industry is no stranger to automation. It is, in fact a pioneer in this area, having started out on its journey in the 1960s. There is very little in the assembly line or supply chain that is not fully optimized, and even less left to gain. From simple mechanization, carmakers progressed to industrial robotics – General Motors introduced Unimate, the world’s first industrial robot – then to digital automation, and today are extensive users of robotic automation. U.S. automakers buy one in two industrial robots sold globally, says the International Federation of Robotics.
A few years ago, Ford introduced a “seeing” robotic arm to install different parts, such as windshield, fenders and doors, on the body of the Ford Escape more accurately. Chrysler’s Sterling Heights Assembly Plant has a robotic flexible body shop.
Handling everything from welding to assembly to painting, robotic automation is indispensable on the automobile shop floor. It has also pervaded car financing and distribution. Going from physical robots to digital bots, taking advantage of robotic process automation (RPA - a term that has become standard across the industry) with bots handling customer queries in the call centers of financing companies, and software robots at dealerships scheduling service appointments, sending alerts, running diagnostics and even selling cars itself. In March 2017, online car retailer Caravana launched its fourth car vending machine in the United States. A buyer can research, purchase and finance a car on the company website and then ask for it to be delivered, or choose to pick it up from the vending machine, which works pretty much like any other. If the buyer needs help, a company rep is at hand.
Data-driven automation of decisions
Digitization brought data and the technology to gather it, manipulate it, analyze it, interact with it and act upon it. Data science has come a long way from business intelligence and early analytics applied to orderly, enterprise owned information to encompass technologies, such as big data and predictive analytics, machine and deep learning, natural language processing and visual recognition, that can handle absolutely enormous quantities of data from a variety of sources and formats, and in varying degree of cleanness. Following from this, the way data is used has undergone a sea change, where it is no longer about studying events past, but about understanding events as they occur in real time, and predicting future events before they unfold. Where data merely provided a diagnosis in hindsight, it is now producing insights that can be used to make smarter decisions and initiate timely actions. All of this can run on its own, without human intervention.
The modern car is a supercomputer on wheels, and its sensors and cameras generate a wealth of data that someday might be worth more than the automobile itself. A sensor fitted on the engine can alert the driver about a part needing replacement; a camera can pick out an empty parking spot. A self-driven car produces about 1 GB of data per second, and it is expected that by 2020, manufacturers will earn more from selling data than cars. Self-driving taxi operators will earn more than cab fare by beaming personalized, location-specific advertisements and promotions. Connected cars will also potentially save manufacturers and insurance companies millions of dollars in prevented accidents.
On the financing side, data is driving automation of several decisions and processes. GM Financial, for example, uses an integrated loan origination solution that can interact with users to process requests or perform transactions automatically. The solution not only automates application processing from end to end – data validation, car valuation, credit scoring, fraud detection, etc. – but also makes lending decisions in conformance to company policies and local and national regulations.
Data-driven automation has also entered the car dealership to impact everything from customer experience to marketing campaign effectiveness and lead conversion. At the Walser Automotive Group, a marketing automation platform picks up CRM systems data to craft individualized marketing messages with content that speaks to the customer’s desires and position in the purchasing life cycle. An example of this is a repeating campaign called “Shoppers”, aimed at customers in the middle of the funnel, which has consistently yielded impressive results.
Intelligent automation of the ecosystem
At the third and highest level, automation will take over the automotive ecosystem itself. Still some years away, the seeds for ecosystem automation are being sown today by technologies such as Artificial Intelligence.
A scenario of that kind could be imagined somewhat like this: A potential customer asks an online chatbot on a car manufacturer’s website or social messaging platform whatever he needs to know about a particular model. The bot transfers that information and context to the car company, which might, using an automated marketing platform, send additional relevant content and promotional offers to the customer and arrange a test drive in a virtual reality environment. Assuming that U.S. car manufacturers are still prohibited from selling new cars online and hence dealerships are still in vogue, the happy customer places an order with the dealer; again a chatbot might facilitate the transaction, including an attractive car loan and a vehicle insurance policy that is crafted based on the buyer’s driving history.
Now the manufacturing plant gets into action, ordering add-ons such as paint and accessories just-in-time from the respective vendors via an automated sourcing platform, putting them on a ready and waiting car body, to be shipped to the dealer for delivery. (Note that this prevents the dealer’s inventory from piling up.)
The customer picks up the automobile, and syncs it with other cars in the family and their mobile phones. He drives to work on his fancy new wheels. Depending on who gets off work first, and the current location, the car advises who should pick up the kids from school. Thirty minutes from home, the customer activates the air conditioning and the oven from the dashboard of the car, and orders a bottle of wine from the nearby connected supermarket which will deliver by drone. Dinner is now waiting.
This scenario can be enlarged in umpteen ways – a connected car that hooks into a ride-sharing ecosystem and doubles as a taxi, or a common pool of cars share-owned by several people, to name just two. Here, a technology such as Blockchain could come into the picture to create transparency and enable a trusted transaction between total strangers.
Undoubtedly, the automobile industry has made tremendous advances in process automation, and considerable progress in data automation, but has barely begun to automate its ecosystem. There is a need to focus attention and resources in this area because it could potentially solve most of the industry’s problems by lowering cost, adding monetization opportunities, reducing unsold inventory, and above all, preparing for the not quite imminent, but certainly inevitable, domination of the autonomous car.
