Thank you for your request!
Please check your inbox in a few minutes for the follow up email from our team.
Smart Car-Sharing
with Shift
Shift, Inc. is a car-sharing service that automates
the car rental experience through an app on your phone.
Start your order
You use the Shift app to create an account, upload a photo of your driver’s license, and select the day, time, and location of your desired pick-up and drop-off. Then use the app to choose a car, select options, and enter your payment method.
Shift shows you
the best matches
The app launches a map on your phone to show you exactly where to pick up your car and how to verify and check out the vehicle — using two-factor authentication for security.
The drop-off process is just as simple, guided by the Shift app.
Customer benefits
People who use Shift to rent cars benefit from competitive prices, the ability to rent by the hour, convenient pick-up and drop-off locations, and no agents, paperwork, or waiting in line.
Corporate clients
Shift’s corporate clients include Amazon, DHL, Honeywell, and Ikea. Shift’s vehicle partners include Chevrolet, Hyundai, Mazda, and Toyota.
Shift landed on Saudi
Shift launched operations in Saudi Arabia, covering 33 cities with 3,500 cars and 1,200 commercial vehicles.
Shift’s goal is to build the world’s foremost integrated mobility platform for people and products to drive business efficiency, quality, and sustainability.
The challenge
To perfect its app, Shift needed to solve two related challenges for artificial intelligence (AI): Arabic optical character recognition (OCR) and interpreting various driver ID documents.
Arabic OCR is still in an early development stage, with limited solutions and no AI training data set. Shift’s app also had to read a wide range of driver ID documents, with various layouts, lettering, colors, and conditions (surface glare, scratches, watermarks, and so on).
After trying existing technologies, Shift concluded that it needed to develop its own advanced AI training model to accurately, reliably recognize and interpret driver ID documents.
Why Diffco
Shift was impressed by Diffco’s technical expertise. Diffco has extensive experience developing apps for computer vision and OCR, including solutions to complex challenges similar to Shift’s. For Shift’s project, Diffco offered a lead software engineer with over 15 years of experience in OCR, including with global OCR pioneers. Shift was also impressed by Diffco’s approach to Agile software development for speed, flexibility, and quality control.
Solution
1
Diffco meticulously collected, prepared, and labeled data to lay the groundwork for AI training.
2
We then developed and implemented a new AI training model, tested it, and compared the results with other approaches and technologies.
3
As we iterated the AI model, we broke the large problems into smaller, discrete tasks and solved them separately.
4
Finally, we combined all the discrete solutions into a single, integrated, comprehensive solution.
The initial AI model that Diffco developed for Shift achieved a recognition success rate of 96%.
Technologies used
OpenCV
Dlib
Firebase ML
Tensorflow
Torch
Diffco also delivered the following for Shift
SDK for mobile devs
SDK for mobile developers that enables easy integration with Shift’s mobile app.
Enhanced enrollment process
Enhanced enrollment process in the Shift UI, using AI recognition to make the user flow smoother and easier.
New server setup
New server infrastructure setup using Amazon Web Services, Docker, and CUDA for improved driver ID document recognition.
Let’s build something
great together
We believe in turning ideas into reality and we're ready to join your journey. Reach out to us and let's start discussing your project.
