Two years ago, Shun Matsuzaka, a digital creative director then based in McCann Worldgroup’s Tokyo office in Japan, constructed the world’s first AI creative director capable of directing a TV ad.
McCann pitted the robot against human creative director Mitsuru Kuramoto to construct an ad, and people had to guess which one was directed by a human. Kuramoto’s treatment narrowly won 54% of the public vote.
Although the experiment showed AI can successfully execute a creative campaign, Matsuzaka says the technology isn’t on the cusp of stealing marcom jobs. Not yet, anyway. He thinks creators should look at AI as a partner rather than a threat.
"So long as we are communicating with humans, there will always be important work for us that can’t be replaced by computers," he says. "At least for a while…"
Experts agree the data AI provides is a source of information that can be used as part of comms efforts and marketing campaigns. The technology enables a creative person to uncover insights they might not otherwise have the skills, resources, or time to discover.
AI’s primary role is to help remove more mundane or repetitive work, enabling marcom pros to free up more time to invest in creative thinking and applying the insights it digs up, says Text100 CEO Aedhmar Hynes.
"Every successful campaign is built on a key insight," she explains. "Getting to that insight requires research and understanding and that’s where AI can most help PR pros with creative - by interpreting vast datasets, building meaning from them, and determining fresh insights."
Alyssa Waxenberg, director of marketing for IBM Watson Marketing, describes the cognitive system’s data insight as a "jumping-off point" for marketing pros.
"They can take that and layer on human emotion or intuition around what is working and bring to life the emotion of a brand," Waxenberg says.
Human emotion, experience, and intuition can’t be recreated by a robot, and those elements are essential for the creative execution of a campaign that aims to properly represent a brand and resonate with consumers, she adds.
"Think about commercials and content - it still needs authenticity of emotion," says Adam Hirsch, global EVP at Edelman Digital. "You need influencers or ambassadors or actors to convey emotion, especially if it’s video-based content. You can’t automate those things."
But AI is already widely being used for the execution of marketing campaigns, says Waxenberg. The technology is helping marketers see what platforms are most successful, and what content is most effective.
IBM customer Performance Bicycle, a specialist bike retailer, is incorporating new cognitive capabilities, including Watson Content Hub, in an intelligent CMS. AI understands and learns about the data in its CMS, then trains itself to recognize and automatically tag content including images, videos, and documents based on millions of previous examples.
The technology tagged content and made connections Performance Bicycle staffers could not have done previously, says Waxenberg.
For example, Watson discovered customers who bought mountain bikes were watching specific videos the team had not previously tapped into, including one featuring desert riding. By making these connections, the retailer can proactively present this content to others in the future to help drive sales.
1-800-Flowers also used AI to better market to consumers. The brand recently took the wraps off an IBM Watson-powered gift concierge called GWYN or "gifts when you need."
Online customers interact with GWYN using natural language such as, "I'm looking for a gift for my mother." GWYN interprets queries and asks qualifying questions about the occasion, sentiment, and who the gift is for to ensure she shares the appropriate, tailored gift suggestions with each customer.
Over time, GWYN learns about customers’ unique gifting needs, enhancing the shopping experience specifically to each individual.
Within PR firms, Shift Communications uses AI for heavy data analysis, insights, and predictive analytics. The firm recently took five years of data about a client and, based on that, used AI to project one year ahead what likely organic search outcomes would be.
"We’re using it to inform editorial calendars, content marketing, and social media [efforts]," says Chris Penn, VP of marketing technology at Shift. "If we know March will be their peak time, we know what short-lead and long-lead pitching we need to do. Instead of being reactive to what’s happening in the industry, we know from predictive analytics what is likely to happen and we can plan way ahead for it."
Charlie Witkowski, SVP, creative director, innovation lead at Olson Engage, says the technology is almost at the point where it can be used to inform a campaign’s strategy and drive its creative direction.
IBM has been studying marketers for two years and training Watson to learn the discipline and how to do it effectively. Waxenberg notes the company is building tech that helps automate routines such as running reports, pulling data from silos, and combining different systems to give brands a 360-degree view of customers.
IBM is focused squarely on helping marketers create effective content that resonates with consumers. Its next AI product is in beta now, but IBM will launch it soon.
"We do email campaigns and know open rates for subject lines and content, but imagine you can do more immediate testing or optimization of creating marketing materials, seeing the best performing visuals, and testing copy," says Waxenberg. "Putting that all together is the area we are exploring."
So while McCann-style Robo-creatives may not yet be fully ready to take over the campaign generation process from their carbon-based masters, the AI underpinning the robots is certainly helping humans tailor effective content and communication that works much harder for brands and organizations.
AI CASE STUDIES
Facing a tough financial decision? Turn to chatbots
Last May, Wells Fargo turned to Facebook Messenger as its primary channel for customer service. This spring, the embattled bank began piloting an AI-driven cus- tomer chat experience on the platform with 5,000 customers and staffers.
The bank’s goal is to use the chatbot to deliver information in the moment to help customers make better-informed financial decisions. It is particularly aimed at reaching millennials and younger, tech- savvy generations.
After a simple registration, customers can ask Wells Fargo’s bot basic questions about their account balance, most recent transactions, and how much they spent on food last week, among other things. Over time, the chatbot becomes more conversational in its responses.
Wells Fargo is the first large U.S. bank to offer a chatbot on Messenger. A bank rep said Wells Fargo is unable to share metrics because the chatbot is still in pilot mode.
Hazelnut spread enlists robot to design its label
Nutella recently set out to "individualize its product through design" and "make each jar as unique and expressive as the Italian people," explains a video from the advertising firm behind the campaign, Ogilvy & Mather Italia.
The brand completely erased its traditional knife spreading Nutella on a piece of bread image for a campaign called Nutella Unica. Using one algorithm, dozens of patterns, and thousands of color combinations, 7 million unique, eye-catching label designs were created. Tech company Hewlett-Packard created the algorithm. The jars went on sale in Italy in February, and every one was snatched up within a month. The brand’s effort garnered 3 million posts on social media, and 10,000 videos were created by people about the one-of-a-kind jars.
Based on the campaign’s success, other European countries will soon be able to get their hands on uniquely labeled jars, according to food community website. Food52.