As the accelerated growth of e-commerce forces online retailers to face new challenges on a regular basis, coming out on top is best achieved via the use of emergent technologies. This can be done, for example, with the use of Artificial Intelligence (AI), which is already showing outstanding benefits in a wide variety of industries. As AI becomes more advanced and ever more sophisticated, the possibilities it offers to e-commerce industries is unparalleled.
One of the ways in which AI can dramatically improve the buying process is by enabling the customer to find the right product in the right moment. This is critical, given that this is the very first step in the online buying process. With this in mind, how can we improve the effectiveness of locating a product in a given retail store? One simple way to make the process easier for the customer is to allow them to search via a photo of a product, rather than having to use specific key words to find what they want. But there’s more to it than simply helping the customer find their product – via the process of using visuals to search, the customer’s buying experience is enhanced overall, thereby fueling their shopping experience.
Although a photo is useful if the customer happens to have one of the item they’re searching for, there are also many more customers who may not have a photo. Or they may have a photo that doesn’t quite match their needs. Here, the alternative is to use a simple sketch or drawing to express what the customer is trying to find. There are a multitude of advantages to using drawing-based querying (a ‘sketch query’) rather than text or a photo – it’s creative, expressive, effective, user-friendly and also fun.
A sketch query is highly expressive – even a simple drawing can convey complex shapes or textures that can barely be described in words. Color is another useful feature that can be incorporated into a sketch.
Computer vision, a branch of AI that is devoted to developing methods of interpreting our world through a set of images, combined with deep-learning-based models, has provided the possibility of finessing several aspects of image searching. Indeed, the computer vision community has become a highly-dynamic field based on the deep-learning models showing tremendous impact in a wide variety of tasks related with visual perception (image recognition, image segmentation, object detection, image captioning, etc.), and allowing us to build ever more effective models for using imagery or sketches to search.
Sketch-based querying is also supported by the technological advances both in mobile devices and in e-commerce in general. Indeed generating a simple drawing using a mobile device is now a simple task, easier than writing out a search query. Since search engines rely on somewhat correct spelling of search terms as well as an understanding of the language, drawing a picture does not. It’s the far simpler and more effective way to find what you’re looking for.
Once the customer has tried drawing something and getting great results from a search, they’ll be back for more. The experience itself is fun and entertaining, and will leave the user with a thirst to try out different sketches, search for even more products and, ultimately, to buy.
In conclusion, using a sketch-based product retrieval method provides a new and innovative way to get the results you want for your customers, reducing the gap between self-expression and search efficacy, and fueling customer engagement.