An image recognition system uses AI to list your car.
Role
UX/UI Designer
Company
Carsales
Project duration
June 2017 (3 weeks)
I. Project Overview
This project began with our GM of AI and Machine Learning creating an image recognition engine called Cyclops.
This innovative AI driven platform designed to simplify the process of selling your car. By leveraging advanced image recognition and machine learning algorithms, this product automates the initial steps of listing a vehicle for sale, making the process more efficient and user friendly.
The Challenge
Build a feature that utilises AI to streamline the car selling process with image recognition and machine learning technologies to automate the initial listing stage of listing your car.
01. Manual entry is tedious and error prone.
02. Users lack technical know how for accurate details.
03. First-time sellers feel overwhelmed.
The Solution
Cyclops uses AI to simplify car selling with image recognition and machine learning, automating the listing process for speed and ease.
01. Image recognition scans the car to extract key details automatically.
02.Auto-fills listing info like make and model.
03.Cuts steps for a smoother experience.
04. Speeds up time to sell.
II. Design Process
Research
We decided to use this tech within the flow of selling your car. The user would have the option to take a photo of their car or skip the the process and add manually.
I collaborated closely with the development team to gain a deep understanding of how the integration user flow would be implemented and executed.
Integration user flow
Defining the Experience
Users simply needed to take a photo of their car using their phone. The AI will analyse the image to identify the make, model, and badge of the vehicle. This feature uses deep learning models to recognise various car parts, logos, and shapes to accurately determine the car’s specifics.
Once the car is identified, Snap n’ Sell automatically populates the listing with basic vehicle information like make, model, year, and sometimes even the trim level or specific features if identifiable from the image.
While the system autofills much of the listing, users can add more details about the car’s history or any upgrades and after market features.
Conceptual wireframes & user flows
III. Final Designs
Snap n’ Sell Screens
IV. Outcome
Snap n’ Sell aimed to revolutionise the car selling experience by combining user friendly technology with sophisticated AI, making the process of selling your car as straightforward as taking a photo.
We also adopted this AI for when searching for cars as away to demonstrate our innovation that was coming out from carsales. If you wanted to know what type of car it was or pricing you just needed to pull out your phone and take a photo.
Home screen search
V. The Impact
Several years later it was redeveloped to cater for the 100,000’s photos a day coming from private sellers and car dealers, no longer needing to manually categorise them all is a huge time and cost saver. Not only that, Cyclops makes possible a smart comparison of images. Users can now visually compare, for example, the backseat photos of a Honda CR-V, Mazda CX-5 and Ford Kuga without having to reorder or sort the images.