IBM and The Weather Company's Cognitive Council Aims to Shape the Future of AI in Advertising

Nearly two dozen agency and brand execs have joined a newly created group to discuss how artificial intelligence will change marketing. Last month, The Weather Company hosted the first meeting of the Cognitive Media Council, which aims to better understand how AI can apply to marketing strategy. The goal: to identify how cognitive computing can be used in advertising and identify key opportunities and challenges associated with using the emerging technology that’s quickly become one of the most hyped areas of modern-day marketing. According to Weather Company CMO Jordan Bitterman, when IBM acquired The Weather Company in January 2016, it came under the umbrella of IBM's Watson division with the mandate to create cognitive solutions. With its Watson supercomputer, IBM has increasingly been varying its use of AI, creating a number of APIs for advertising and matching social influencers with brands. (In September, The Weather Company launched Watson Ads, with brands like Campbell’s using it to create ads in the form of personalized recipes based on a users' ingredients and the weather where they are.) It's also showing how it can branch out into other intersections of art and science, such as when Watson ingested the works of Spanish architect Antonin Gaudí to help design a sculpture at this year's Mobile World Congress tech conference in Barcelona. The Weather Company selected senior executives from nearly two dozen brands and agencies that have shown an interest in using emerging technology as a part of their marketing strategy. Attendees included representatives from brands like Chase Bank and GlaxoSmithKline and agencies including Leo Burnett. Others in attendance represented retailers and consumer-packaged-goods brands. [kickout id="1"] According to Leo Burnett North America CEO Andrew Swinand, who attended the meeting, data is "like soup du jour," explaining that all of the brands are talking about it even if they're not actually using it. Swinand mentioned a survey he read a few years ago that found more marketers are prioritizing data even if few are using it to make decisions. However, he said, clients are "starving" for it, both because they want it and because they know they need it. But data is overwhelming. "Right now in the marketplace, there is just this overwhelming amount of noise," Swinand told Adweek. "If I told you it's 38 degrees in Chicago today, interesting, but is it relevant for you? And what are you going to do with that? And I think right now in the market, there is a lot of irrelevant data in the market being shared that's not actually important in decision making. What's really interesting with AI is real-time decision making informed by data." Swinand, who joined Leo Burnett in January, said one thing that attracted him to the role was that it's easier for agencies like his to learn to ingest data than it is for data companies to learn to be creative. “I feel like the battleground that’s about to occur is who can best apply data to creativity that has the power to transform human behavior," he said. Susan Canavari, chief brand officer at Chase Bank, also attended the first meeting. While Chase hasn't used AI yet, she told Adweek the company understands that "AI is going to change everything." She also pointed out that customer expectations are changing quickly. "I think the hard thing for all companies is marrying customer expectations of what we should and do know about them with the experience we give them in a way that protects their privacy and their data," Canavari said. Like all of digital marketing these days, at the center of it all is data, and Watson uses both structured and unstructured data to learn in a human way but at computer speeds. Bitterman said AI can help programmatic advertising get smarter as it learns without compromising first- and third-party data from various brands. It could also help marketers "think about data differently," Bitterman said. "Brands, agencies, media companies—we've all become pretty adept at data, but this throws a whole new wrinkle into things," he said. "And what it might do is, I hope, create a filter for thinking about data differently. And if we do that and create different types of filters to use data differently than we did before—the structured and the unstructured—then everything can get digitized. And if there's a system that starts to learn and reason, it becomes more user-friendly."