We’re currently in New York City to attend our second National Retailer Federation ‘Big Show’. In true American style – there are thousands of attendees and the worlds best retailers espousing their wisdom.
So, if last year’s programme was all about omni and the usual ‘the customer doesn’t think about the channel just the brand’; this year has been a huge jump forward to AI (Artificial Intelligence). This is a topic being grabbled with in tech-based industries – for example, Telsa’s CEO Elon Musk is famously cautious about humans venturing and exploiting AI.
I’m not going to get into the depth of AI, but rather one application – North Face and IBM’s Watson.
Watson famously competed as a contestant on American Game show Jeopardy. Now it’s intelligence has been used by North Face to product an AI personal shopping application. This is limited to jackets and to their website, but allows online customers to find exactly what they’re looking for by filing in a couple of questions – in a way that you would normally talk. For example, if you don’t know something you can just type that in and the programme won’t say ‘error’ – like the usual decision tree problem solvers we’re all used to.
The questions go along the lines of:
1. Where and when will you be using this jacket (set context for choice)?
2. Is it for a male or women?
3. What activity or sport will you wear this for?
4. What colour jacket do you want?
5. Do you need a heavy, mid, or light jacket.
6. Do you want a down or synthetic insulated jacket?
In it’s early stages the results have been successful. According to Senior Director of eCommerce, Cal Bouchard, the tool has had:
  • 50k used the tool in 60 days and 2 mins average session;
  • 2.5/3 rating for usability;
  • 75% said they would use XPS again;
  • Maybe most importantly, people with the most trouble loved.
What I found most interesting what they lessons that the North Face team had learned through the process of working with AI, they were:
1. There’s lots to teach: AI seems like it should be easy – not true. Everything that’s obvious you have to teach the system to know. For example, you have to teach it that running the same thing as jogging. This you take for granted.
2. You need lots of partners: this who will help you teach. Those with deep knowledge into product DNA and catalogue. Ask the question your customers want answers to.
You should definitely check out the functionality on the website below – it’s a beautiful step into AI world.