Imagery, and the conversational interface.
When the Rosetta Stone was first uncovered in 1799 a window to understand antiquity was opened. There was finally an index of terms to compare to the hieroglyphs of Ancient Egypt and as a result, it was finally possible to understand the pictographs they left behind.
Combined, computer vision and deep learning technology could be called the CPU’s Rosetta Stone. The human affinity toward images combined with the ubiquity of social media has shifted the paradigm of what it means to communicate. Deloitte announced that in 2016 3.5 million photos were shared every minute, with 90% being taken with a smartphone. A similar report by DOMO saw that number grow to 3.7 million in 2017. Until recently the capacity of indexing and classifying these images was limited to the metadata supplied, the alt tag, location information, date captured etc.. but when it came to the pictures, computer’s saw a black box.
In the last few years consumers have started interacting with computer vision powered interfaces in the wild. Examples like facial recognition to auto tag photos have made consumers expect smarter experiences and the AI models driving them behind the scenes have become more capable and accurate. Niche AI companies, like ScopeMedia, are working to bring targeted learning & expertise to select industrial applications.
ScopeMedia has developed an experience to act like a personal stylist intended for use in fashion retail. As we developed our computer vision technology we noticed that teaching a computer to see gives it a window to the quirks that make up our personal preferences. The visual information categorized by our models proved to be a massive insight for the computer when searching for relevant products in large inventories.
Customers experiencing the interaction have been observed as having a conversational flow, echoing a pattern of prompt and response. They are able to communicate their preferences to the stylist through their common language of imagery, and as their conversation progresses the AI will get a better understanding of a customers personal taste and in turn present more relevant options.
You tell a computer what to do – You initiate the search – You open the app. Whereas experiences like Stylist are different, they speak first and you to respond. By allowing devices to guide users to their first step in a journey helps eradicate any cognitive load required for an initial interaction. In essence, systems that present personally relevant options to customers allow for more fluid interactions with technology.
Stylist allows static spaces to be reactive, predictive and personal and there are more examples of similar interfaces popping up in the wild. Before we know it, it will be the norm to expect similar touchpoints global brands and institutions.
So, where does this leave us? The retail landscape is at a tipping point, some of the systems that power commerce today are evolving. Their becoming more fluid to interact with and better at predicting customer expectations. By training interfaces to recognise the context of the visual world we give them the capacity not to understand, but like the rosetta stone, index our visual experience and as more services start to adopt this style of conversational interface our interactions with previously static spaces will inevitably become tailored and personalized.