If online retailers want to make more sales and retain customers, they better stop making those clients do all the leg work. Randy Dean, chief business officer at Sentient Technologies, urged the audience at WIRED Retail to stop expecting customers to use outdated interfaces that are essentially databases asking users to search, browse and narrow the field by selecting multiple options. Our internet connections have become infinitely faster over the past decade, yet retailers still expect customer behaviour to remain the same.
“We need to change the way people interact with our products. We have to get rid of the filtering systems and search boxes, and let the products speak for themselves.” AI, he says, can give products that voice. “We don’t want to live in a world confined by metadata and tagging. We need intelligent systems that can understand nuance.”
At Sentient Technologies, that demand has led to the creation of Sentient Aware, an AI that understands products at “a very granular level, visually”.
Pointing to the example of a shoe, Dean said: “AI is breaking it down into hundreds of vectors of analysis - the shape of the toe, the height of the heel, embellishments; all the subtle things that attract us to the products we buy. It, and we, can understand those things in real time.”
Dean claims that after a view taps, the AI can understand what it is the customer wants and be able to make appropriate suggestions that lead to sales. The technique can be used with any product that has subtle visual differences the AI can tap into, from glasses to shoes or furniture.
The dramatic shift this could lead to is greatly needed. Worldwide today, he said, the e-commerce conversion rate for websites is just 3 per cent - incredibly, retail websites have a 97 per cent failure rate at engaging customers. “Even when they come to our front door, even when they add products to the cart, we are losing the vast majority of customers.”
By using complex neural networks, the technology can constantly optimise a website in response to a consumer’s behaviour. “The site will evolve in order to get the best algorithm.”
Instead of AB testing, the system will then allow brands to deploy thousands of tests on thousands of ideas at the same time. “You can let the AI work out what works and what doesn’t. We are not seeing base improvements, we are seeing double- and sometimes triple-digit improvements in engagement and conversion of customers, whether that’s buying something, putting a name on a list, or clicking on an article, we can optimise towards it.”
Dean believes, ultimately, that changes to the very DNA of a website will start converting that 97 per cent failure rate into better news. And AI will be at the heart of those changes.
“We talk a lot about AI in the future, that AI is coming. But AI is here, you interact with it every day. And given the lesson of history, the sooner you embrace it the more likely you are to succeed.”
This article was originally published by WIRED UK
