August 15, 2024 | By Michael Bürgi
Originally Published on Digiday
A recent study that aims to shed light on AI and its application in marketing and agency work reveals that the more you know about AI, the less you’re convinced it’s going to replace human beings.
That’s one of the findings of a joint study between Omnicom Media Group agency PHD Worldwide and WARC, which surveyed 700 marketing and agency executives (in 10 different countries, including the U.S., U.K., Australia, China, Brazil, and others between May and June of this year) to better understand their plans for AI’s incorporation into their work between now and the end of the decade.
The study fits into a broader initiative by the media agency to become a hub of learning and experimentation around generative AI that’s accessible to all Omnicom clients (not just PHD or OMG) as well as the general public.
To that end, PHD this week is launching what it’s calling a “generative publishing platform” that’s AI-assisted, called Ascension, according to the media agency’s global CEO Guy Marks and global head of strategy Mark Holden. The cost, said Holden, is “well into the six figures,” but he declined to be more specific.
Marks, who said he wants to ensure the Ascension launch isn’t seen as another headline-grabbing bit of bluster or BS, described what he wants clients and the public to get out of it.
“It’s essentially got two speeds to it,” said Marks. “How do we architect that narrative and guide clients and our people — and more widely the industry — about our view of where we are headed and how everyone needs to prepare for that? But also, what do we do in the next 30, 60, 90 days, and how do we prepare not only our own talents, but our client organizations, to leverage the benefit that’s now in front of us?”
Holden likened the advent of generative AI to a “kind of Cambrian explosion moment — everything changes,” he said. “Our point of view is that if you’re a an agency or any form of a supplier, in a marketplace where there’s such rapid change that impacts your clients and of course your employees, in a sense, it’s your duty to put information out to educate and keep people up to speed. If you don’t, then people don’t keep track of developments, and therefore the connection between you and your clients [is lost].”
The study with WARC helped to inform the editorial mission and POV of the “magazine,” which breaks down the rest of the decade into three eras of AI evolution.
- The AI Experimentation Era (2024-2026), which will be marked by the individual use of large language and diffusion models (which Ascension uses), and the early integration of generative AI into enterprise platforms. It will be a period of pilot projects and refinement, ushering in the next stage.
- The AI Acceleration Era (2026-2028), during which AI “becomes a fundamental element of marketing – with more mature applications of generative AI in existing platforms. It will be used by a large portion of the workforce, albeit mainly for efficiency and effectiveness,” according to Ascension. Focusing on replacing and/or enhancing current functions, this era will lead to…
- The AI Elevation Era (2028-2030), which will focus on more intelligent systems that can make more reliable decisions, freeing workers dream up “new marketing capabilities beyond what we can envision today,” according to the study.
Greg Stuart, CEO of MMA Global, an industry organization that’s diving deeply into AI experimentation and testing for its marketer and ad-tech members, said he feels agency efforts around AI and new tech developments often appear deep but don’t scratch much beneath the surface.
“The question I would ask is, what are the frameworks they’re advising for clients to look for opportunity within AI?” asked Stuart. “What are the opportunities to frame where marketers need to dig in? And what knowledge do they have now that nobody else does? If [agencies] can’t answer those questions, it can appear superficial.”
What did the study find?
Among the highlights are a gap between the desire to adopt generative AI and how much it’s actually being used. While 35% of marketers feel it should be used to a high or extremely high extent, only 27% are using it at those levels. Agency side respondents were at 36% and 26% respectively.
The study also quizzed the respondents to gauge what they said they knew from what they actually knew. While 42% of marketers consider their generative AI knowledge advanced, only 13.7% of all respondents correctly answered two or more out of five quiz questions.
Finally, as knowledge bases increase in understanding generative AI, the more respondents understood that human participation will remain necessary for success. Only 21% of higher knowledge respondents (i.e. those who answered at least two of the five quiz questions) believed that generative AI would replace actioning tasks currently carried out by agencies in the short term, versus 46% for those with below-average knowledge.
“The more you know about gen AI, the more you go, ‘hang on a minute,” said Holden. “There is no big green button … The whole thing just requires a displacement of roles into different functions, as opposed to replacing” the functions.
MMA’s Stuart said PHD/WARC’s study results illustrate the gap that will soon separate the marketers who figure out AI faster than others. “You must figure this out fast, because it is going to steamroll us all,” said Stuart. “It’s a little bit like seeing a tornado way off in the distance and deciding to start working to shore up safety at that moment — when you should have been working on that two three years ago.”
