What to expect from AI and Visual search in 2018
Artificial Intelligence (AI) is geek speak for getting computers to do a ton of highly complex research into customer behaviour or database interrogation in the blink of an eye and present some interesting results. If you’ve ever used a SatNav you’ll be familiar with how AI works as your SatNav identifies the best route to your chosen destination, takes into account the time of day and any traffic incidents and guides you to wherever you wish to go.
For retail AI becomes really interesting as it sits behind product recommendation engines, search engines and powers new forms of search such as voice and image based input. Voice search is markedly different as it’s generally natural language interspersed sometimes with “Please” and “Thank you” as politeness kicks in when speaking compared to text input which tends to be less verbose and restricted to a string of key words.
Perhaps one of the most interesting applications of AI in retail is Image Search. Consumers often struggle to describe the things they wish to buy but sharing a picture, either taken on a smartphone or tablet camera or simply shared from the web, is much easier. It’s a ‘See it buy it’ mentality which is simple for the consumer but a much more complex proposition for the retailer.
Some retailers have already invested in image search. Notably John Lewis became the first UK retailer to offer visual search functionality when Cortexica technology was used to power the ‘Find Simliar’ tool in their iPad application. Initial tests commenced in January 2016 with expansion to more categories in July of the same year.
“Find Similar is an online function that displays search results of a range of products within a particular category that are visually similar in colours, shapes, details and patterns. By searching visually through the inventory, the customer has access to a greater range of product choice and inspiration without the need for inputting keywords.”
– Cortexica describing John Lewis’ iPad app
It’s not just streamlining search on a site where visual search can benefit, if a product is out of stock AI could potentially automatically search the image database for similar products rather than simply return a null search result. Cross selling opportunities also exist, for instance displaying accessories alongside an outfit could lead to an increase in basket size and “Like this see similar” links can increase dwell time on a retailers marketplace or website.
Visual search is approaching a tipping point where it will become mainstream and, perhaps not yet on desktop buy certainly on mobile, a commonplace method of searching. It’s a technology that lends naturally lends itself to certain categories such as fashion and with the ever decreasing cost of computing power and ready availability of AI technology 2018 will likely be the year of wider adoption.