The technology can enhance creativity by taking over marketing's more mundane tasks. But will AI ultimately replace human beings?
An artificial intelligence engine just bid on an ad while another cut a movie trailer. Will the technology be coming for your job next?
The question of whether marketing is more science or art has never seemed more relevant now that highly sophisticated cognitive learning technology is able to assume many of the tasks involved in marketing — in some cases, even doing them better than a human could.
But visions of a completely automated campaign may be premature, according to executives from IBM and other companies at the forefront of AI who weighed in on the technology’s impact during a panel discussion at ad:tech New York last week. In good news for creative directors, the experts said cognitive technology has the ability to free up marketers to spend more time tackling bigger picture responsibilities, such as finding the inspiration for the right voice and vision to make an emotional connection with consumers.
By laying the groundwork for significantly more sophisticated one-to-one marketing, AI could even create a need to beef up analytics, content and other areas for businesses that are able to gain a competitive edge through customer-centric marketing.
“Right now, consumers have figured out that ads are fake and experiences are real,” Jason Alan Snyder, chief technology officer at Momentum Worldwide, said during last week’s ad:tech panel. “Advertisers are making a toxic products. It is very helpful to have something that can make inferences — my job is to create intimacy at scale and AI helps me do that.”
Here's a look at how AI is learning to be a marketer and why the industry may never be the same again.
The first step
There are two visions of AI: One is of a futuristic world where a single, powerful intelligence exists that people can connect to no matter where they are. This Star Trek-like version may be coming down the pike, but it is not likely to arrive for many years, according to the experts who spoke on last week’s panel.
The second version — what we refer to as AI today —is really more of an augmented intelligence, explained Ari Sheinkin, VP of marketing analytics at IBM. In this version, cognitive technology — by virtue of repetitive learning — is able to take on some tasks traditionally performed by humans.
Already, examples are piling up of how AI is impacting the consumer experience. Apple's Siri and Amazon's Alexa use natural language processing to enable users to make a request, such as finding a song online or making a purchase. An IBM Watson-powered ad for Campbell’s on The Weather Channel asked visitors what they had in their cupboards and recommended a dinner recipe, while a similar deployment for North Face queries customers with questions to help them find the right item.
“It becomes a really human moment for people,” said Sheinkin.
The experience is not seamless, however. There are numerous reports that, in their current form, chatbots have too many limitations for broad appeal.
“It is the first step in a journey,” said Dave O’Flanagan, co-founder and CEO of customer intelligence cloud software company Boxever. “When they become more intelligent, they are going to become smart and they are going to do it quickly.”
Crunching the data
With marketers struggling to leverage the wealth of data available to them these days, it is perhaps no surprise that AI’s next frontier appears to be behind the scenes, supporting a variety of tasks involved in the mechanics of marketing.
For example, one of the trailers for science-fiction horror film Morgan was cut by IBM Watson, the cognitive computing system that's able to watch content and interpret the language and emotional tone of each scene.
IBM is also leveraging AI for programmatic bidding. With programmatic, marketers have access to a significant amount of information about target consumers and must make lightning-quick decisions about whether or not to place a bid. AI can make the necessary computations more quickly than a person while also learning over time what the most effective moves are. Already, the AI strategy is showing a 35% lift in results during early tests.
AI can also learn to tag content, something that is a crucial — but not very exciting — task as marketers increasingly look to develop value-add services, like Spotify recommending songs for listeners.
AI is even impacting how creative is developed.
Earlier this year, agency McCann Erickson Japan appointed an AI creative director and had it develop a campaign for Clorets mints. A second campaign was created by a human creative director. When consumers were asked to vote online for their favorite, they chose the latter, but the results were close.
These results suggest cognitive technology cannot yet replicate the inspiration and emotional connection required for a truly standout campaign.
McCann also tested AI’s ability to optimize the creative for an outdoor billboard promoting a fake coffee company. A camera above the billboard watched how consumers reacted to different creative and copy, evolving the message as it homed in on which elements generated positive reactions.
McCann is even helping consumer packaged goods companies with their sampling and trial efforts using IBM technology that provides block-by-block insights into groups of consumers and adapts in real-time to what is happening on a block to uncover shopping marketing opportunities.
“We are talking about helping creative people move more quickly,” said IBM's Sheinkin.
Finishing the story
Looking ahead, Momentum Worldwide, Boxever and others expect to leverage AI to automate the creative process. They imagine creative directors will develop multiple headlines, taglines and other creative elements for a campaign — all of which are then interchangeable. An AI engine would combine these elements in any number of permutations and learn which versions perform best to continually optimize a campaign.
“The role of the creative is that they are generating stories,” Momentum Worldwide's Snyder said. “Moving forward, the way we are trying to leverage that is to generate stories that are incomplete. Then AI can provide tools for completing the story. This is a significant shift in how we leverage these technologies.”
With AI able to play so many different roles in marketing, it's not surprising that some marketers are scared their jobs are becoming obsolete. But IBM’s Sheinkin believes they have nothing to worry about.
“When we roll out a new capability, our experience is the exact opposite of what you see in the press,” he said. “We go out and hire new people because our capability is so much more. I think you will see more of that.
“In a world where everyone is becoming more data dependent — what is the differentiation? We are moving toward a cognitive world but it doesn’t mean Watson is going to replace jobs.”