IBM's Watson aims to pick the fashion trends

Summary:IBM is working with retailers to get the company's cognitive computing system Watson to help customers decide what to wear.

IBM's Watson will use data from social media feeds, advice from friends, moods, and previous outfit decisions to help people choose what clothes to wear, according to IBM's global retail transformation leader for Watson, Keith Mercier.

keithmercier
Image: IBM

IBM is seeking to commercialise its Watson cognitive computing platform outside of its traditional areas of healthcare and medical research into new fields such as retail and call centres.

In July, IBM and Apple signed a deal that would see the two companies collaborate on a number of apps targeting retail, healthcare, banking, travel, transportation, telecommunications, and insurance industries.

Mercier, who was in Australia this week for Melbourne Spring Fashion Week, spoke at an industry event on Monday designed to encourage retailers to begin looking at how mobile applications could use Watson to assist customers in choosing clothes.

The company released the results of a survey in early August indicating that customers in Australia are now expecting to deal with retailers both online and in-store, and retailers should be able to transfer interactions with customers from online to in-store, and vice versa.

Mercier told ZDNet that an ideal application that retailers would use would allow a customer to tell Watson what sort of mood they are in, and Watson would be able to choose an outfit based on that information, along with other data pulled from social media and other sources.

"If you know the places where consumers are getting data today, when it comes to shopping, it is weather, it is inspiration from social, it might be what is in my closet, it might be my purchase history. You could use Watson in other technologies to bring all those together, and then have that dialogue around the data," he said.

But Watson wouldn't be the fashion guru making the final decision on what you wear, Mercier said.

"When I say Watson gives you an answer, I mean it in terms of advising you. Ultimately, you are the one to decide what you wear."

He said that when IBM works with retailers on apps using Watson, the company seeks to discover what data the retailers are trying to use, and then what questions they would want asked in order to make the best use of that data in Watson.

"Only then do we start to plug in and then train Watson on the types of questions, and the types of answers we would expect. It doesn't have a mind of its own; it's a closed-loop system that way. It only knows what you teach it," he said.

For retailers that may deliberately not give Watson all the information to present to customers, Mercier said customers would ultimately know better.

"Businesses are going to do business the way they want. I think what we're seeing is that consumers are smart. I would hope the best implementation [of Watson] would be an authentic one, where you're using Watson to pool in not only what we want you to hear from us, but potentially what you might go out and do on your own anyway," he said.

"So maybe, if we support that behaviour because we're giving you the capability, it might actually build more trust for you in our brand.

"But you could train Watson to do what you want to do, and set business rules around what you push; that's no different from what businesses do today."

Josh Taylor travelled to Melbourne as a guest of IBM.

Topics: IBM, Australia, Big Data, E-Commerce

About

Armed with a degree in Computer Science and a Masters in Journalism, Josh keeps a close eye on the telecommunications industry, the National Broadband Network, and all the goings on in government IT.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Related Stories

The best of ZDNet, delivered

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
Subscription failed.