Artificial Intelligence is proving popular with North Face customers
Outdoors specialist The North Face is utilising Artificial Intelligence (AI) to improve the customer experience online by helping shoppers find the right products on its website. By Glynn Davis in New York
Speaking at Retail’s Big Show 2016 organised by NRF in New York, Cal Bouchard, director of e-commerce at The North Face, told delegates that search and filtering tools were not the most effective way for its customers to find the right products for their specific purposes.
“We wanted it to be more personal and intuitive with natural language being used so the customer could say, ‘I want a jacket for a skiing trip'” she says. The solution was a partnership with Fluid Group to use the IBM Watson AI technology solution that began early last year and is now delivering positive results since its launch on The North Face website in mid-December.
By asking the customer a series of questions, it gradually hones in on the ideal product for them – determined by their specific needs that are worked out from the intelligent questions Watson asks.
Throughout the interaction with the customer it throws up a selection of potentially suitable products that then give visual cues to the shopper, thereby helping the solution move closer to the ideal item. “Visual engagement helps and the results change dramatically as more questions are asked. And customers can ask questions, which would not be possible with filters,” explains Bouchard.
As the potential products are selected, Watson gives the reasons why it has picked out these specific items. It was initially tested on The North Face VIP loyalty programme customers. Of the 50,000 people that used it the average time spent on it was two minutes and 75% stated they would use it again.
“The customers who would gain most value are those who don’t know about the product they want. It looks super promising to us after 12 months of effort,” says Bouchard, but she warns other retailers who are thinking of using AI that they must take into account the work load involved in creating the data structure that enables Watson to be fully effective.
