Uber and Lyft Are Taking Artificial Intelligence Along for the Ride
Part 1: How Uber Uses AI to Improve Rider Experience
Uber and Lyft are going head-to-head to make artificial intelligence work for them … but in different ways. Which tactics will win? Let’s take a look at where each company is focusing and how they are using AI to get the better of their competitor.
Uber’s Machine Learning System
Where will taxi companies, such as Uber, go next with the help of artificial intelligence? Today, Uber’s primary focus has been on enhancing their profitability and ease-of-use when users book a ride. Uber tracks millions of metrics each day to perfect their services, company-wide.
The first experience with Uber is through their customer portal, and their goal is to make it a reliable experience for riders. If they didn’t get ride time estimates and rates correct, they wouldn’t have a customer base for very long no matter how inexpensive or convenient.
How Might VR Change How Uber Works?
Ideally, with the enhancement of virtual reality, riders could be able to view where their car will pull up and see the vehicle’s make/model beforehand to avoid vehicle confusion. It could even assist drivers in the drop-off visualization, based on congestion in real time.
Let’s say other companies wanted to advertise during rides, as well. For example, Disney could add virtual characters to join you on your trip and converse with about their new movie through a special Uber app. As Uber looks to expand their offerings and market share, they will be looking for creative ways to use these trips for more advertisement opportunities.
Data Collection and Deep Learning
Ride sharing, fake account identification, pickup locations, and UberEATS time predictions all contribute to the vast data being collected by Uber. As more drivers are added to the company, so does the amount of data collected for routes, timing, and (ultimately) self-driving cars. More information gathering means more precise routing, fare estimates, and mapping. Traffic, time of date, and congestion are all taken into consideration for an easier ride.
For these types of predictive models and probability creations, Uber has created Destination Prediction, Michelangelo, and the Bayesian neural network (BNN) architecture. These programs use data monitoring and deep learning to allow massive data to be accessed and responded to, evolving over time to create more intelligent actions. In effect, they help the company “see into the future” with precision.
One area that is being perfected by these programs in particular is the volatility of rider needs during holidays. Artificial intelligence allows Uber to more accurately forecast supply and demand during these high-use times and to detect problems before they happen.
Uber also uses historical data to match riders with drivers more accurately. The goal is to create matches related to distance, time, traffic, direction, and driver/rider desire. They set up budget planning and resource allocation, taking into account weather, population, and other factors.
UberEATS Is Intelligent Too
One of the key ways Uber has expanded their market offerings is through their revolutionary food delivery program. UberEATS’s key difference from other carry-out delivery apps? The Michelangelo artificial intelligence program that allows users at the company to interact with data, create new components, and use models to solve problems.
The next step for AI in this space is to help create a lasting experience with customers. Virtual reality visualizations could accompany special types of orders to include a special themed game play. For example, if one ordered Chinese food, there could be a game/video related to the soon to be released Kung Fu Panda movie.
With more people from the next generations valuing experiences over simple purchases, virtual reality might be a way to offer promoted stories, games, or a celebrity endorsements to sell related products.
Uber Self-Driving Cars
Uber’s next initiative, much like many of Silicon Valley’s giants, is self-driving cars. The most widely accepted self-driving tech costs about $80,000 to outfit one car. The cost for AI in self driving is very expensive, but as more companies pour resources into it, the more it will drive the price tag down.
One of the main opportunities to build rider trust maybe to have Siri type robots speaking during the car ride on the directions, traffic, speed, and safety calls if there are any problems. This would be an easy addition that would help to further differentiate their services as they dive into driverless cars.
More About This Topic
Come back for the other parts of this series, as we continue to explore how Uber and Lyft use AI.