Auto Navigator: Disrupting Car Buying for the Sake of the Customer

By Sanjiv Yajnik, President of Financial Services at Capital One.

At the 2018 SXSW Interactive Festival, we will demonstrate how we are reimagining the car-buying process using human-centered design, forward-thinking, immersive technology, deep auto industry expertise, and innovative partnerships.

The Obstacle

Do you recall how long you spent in the dealership when you purchased your last vehicle? With keys in hand, did you leave the store feeling completely confident in your purchase?

According to a recent survey commissioned by Capital One, fifty percent of respondents indicate that researching and purchasing a vehicle takes more time than selecting a college and selecting a baby’s name. In addition, 62% of car buyers are uncertain that they received a great deal the last time they purchased a vehicle. In addition, 78% of Americans admit they lost confidence in their ability to obtain the car they desired during the shopping process the last time they purchased a vehicle.

We pondered why something as exciting as purchasing a new vehicle necessitated such anxiety. As we delved deeper into the car-buying process, we came to realize that some people are left disheartened when they realize the payments are (much) higher than they anticipated because the total cost of ownership exceeds the price of the vehicle. However, they frequently do not know this at the beginning of the shopping process, when they have their hearts set on a specific automobile. What then? Should we return to the drawing board? Do they begin again, or do they settle for something that exceeds their budget?

The Treatment

In 2015, Capital One announced the first iteration of Auto Navigator, a cloud-based online application built on a microservices architecture and driven by a portfolio of technologies, including machine learning, in response to these client pain points.

This innovation streamlined the car-buying process for clients, allowing them to locate, finance, and acquire their next vehicle with ease, convenience, and assurance.

With our second version, launched just one year later in 2016, clients were able to view more than three million cars from over 12,000 participating dealers around the nation! Utilizing real-time data, we were able to predict how much clients would spend for any car in our database, based on the options they select (e.g., cash down, terms, etc.). In addition, buyers can now pre-qualify for financing without affecting their credit score prior to entering a dealership.

We have iterated on Auto Navigator from its conception, employing human-centered design where we use empathy listening, observation methodologies, and rapid prototyping with customers to continuously create, fail forward and quickly, and improve our customer experience.

What’s Next on This Journey of Disruption?

We are in business to compete with the market for the benefit of our customers. At SXSW, we are showcasing the most recent iteration of Auto Navigator in the form of a new feature:

• Augmented Reality Feature:– We have recognized that automotive research is an ongoing activity that buyers engage in even when they are not actively looking to purchase a vehicle. In the next months, we expect to provide our clients with an augmented reality experience through the Capital One Mobile app, enabling them to scan cars using their mobile devices and receive customized information about each car. This might contain their pre-approved financing, their projected monthly cost, local dealerships where they could purchase the vehicle, and much more. Imagine if customers had real-time access to this type of customized information at the point of need!

Technology Behind It

Upon debut, the Augmented Reality feature will recognize and price cars using Apple’s recently-announced augmented reality framework (ARKit) and their new mobile machine learning framework (CoreML), as well as Capital One’s proprietary models. Using ARCore and Tensorflow mobile, this capability is also compatible with Android.

CoreML allows us to take images from the video stream and feed them through our Convolutional Neural Network to make predictions about the car, while ARKit enables us to recognize surfaces and planes and generate a 3D scene that we can then use to place our AR items. We’ve enhanced the user experience so that consumers may scan numerous automobiles in a single session and obtain additional information about a car they’ve already scanned (even after scanning several). Then, we use Capital One’s proprietary models to personalize the offer so that the customer may view their expected monthly payment.

Empowering people to feel secure about their connection with their money is at the core of what we do, and with Auto Navigator, we’ve found a way to do that while also giving the ideal digital experience that integrates more seamlessly into people’s lives.

These are the opinions of the author. Unless otherwise specified, Capital One is not affiliated with or endorsed by any of the firms mentioned in this piece. All trademarks and other forms of intellectual property displayed or used belong to their respective owners. This content is the property of Capital One.