Sramana Mitra: Interesting. Would you like to do other use cases?
Josh Sutton: Yes, so that was all around insight generation from a structured data point of view. Next I want to move to insight generation from unstructured data. This is an area that I think has massive opportunity. Most of the information that people share is not in a structured format. It’s in the form of call center conversations. It’s very rich but very difficult for traditional platforms to take advantage of, and also very difficult for machine learning models to capitalize on.
This is an area where we’ve been using some natural language understanding capabilities from different platforms to do two different things. One is for a specific brand, understand what is being said by their customers to help steer marketing decisions. It’s basically turning the real world into an active focus group. You’re able to take that real information and pull concepts out of it, the same way you would have somebody do in a structured fashion.
Sramana Mitra: Where are you drawing the data from? Social media conversations?
Josh Sutton: A wide range – from social media conversations, transcripts of call center logs, call center conversations, chat forms. Basically, any type of customer interaction that provides the ability to have a conversational trail. There’re a number of platforms out there that you can do with that. The platform that I think is best is Luminoso, which came out of the MIT Media Lab. From a natural language point of view, they’re blowing everybody away right now. They’re doing great.
The other area around this that’s interesting is some work that we did for Dove. This is in using the voice of consumers to become part of your advertising campaign and not just making it to push message but really using AI technology to, systemically, mine what the collective universe is saying in close to real-time so that can become the story. You can take what people are saying and use that to further your message.
What we did with Dove was we worked with them around a major sporting event that happens in every four years. I don’t think we’re officially allowed to say it out loud for whatever reason. They wanted to do a campaign around female athletes and positioning all the great stories around what people were saying about female empowerment.
We use natural language understanding technologies to mine everything that was being put out there from a journalistic point of view and from a social media point of view, and refine that down so that this campaign can become a breathing and living thing that was using real stories told by other stories to further that brand story that Dove wanted to tell. We’re able to generate 1.3 billion PR impressions. We got impactful results from that. That was around creating a world where beauty is more a source of confidence and not anxiety.
Sramana Mitra: What is your experience? Do you have any clients where you’re using Watson extensively?
Josh Sutton: That story I gave was a Watson example. I’m actually going to be on stage at the Amplify conference talking about that example.
Sramana Mitra: Talk to me a bit about your experience with Watson. You gave me a very good overview of your experience with Azure. Talk to me about Watson. What is it like to work with Watson?
Josh Sutton: Watson has the broadest breadth of capabilities within the platform right now. Potentially, that is its greatest strength and weakness. There’s a tremendous amount of capability within the umbrella that is Watson. As you’re probably aware, Watson is not a single product or even platform, but a wide collection of different services.
David is working to make that a cleaner and easier experience for people to be able to leverage. The upside of Watson is the breadth of it and the commitment that it has from the IBM organization. The challenges that it has right now is that it’s still so broad and they’ve grown so fast that, oftentimes, there’s confusion on how you put the pieces together. We’ve made a lot of strides in working with them to help to figure out how to accelerate that and maximize the efficiency of that.
This segment is part 4 in the series : Thought Leaders in Artificial Intelligence: Josh Sutton, Data & Artificial Intelligence Global Head at Publicis.Sapient
1 2 3 4 5 6 7