Aug 23, 2021 - Technology

The AI adman

Image of robot with Buy
Illustration: Aïda Amer/Axios

Marketing and advertising companies are increasingly using AI models to track trends and generate slogans.

The big picture: Marketers and advertisers focus on two things: identifying and predicting trends that indicate what consumers want, and shaping messages that will appeal to them.

  • AI — and especially new language models that can be tuned to produce messages far quicker than even the slickest "Mad Men" copywriter — are ideal for both.

By the numbers: The global market for AI in advertising and marketing is valued at more than $12 billion, and it's projected to reach $107 billion by 2028, according to a recent report.

  • Research by Gartner predicts that by 2022, more than 30% of digital content will be created with the help of AI.

How it works: AI models are particularly good at drawing in vast amounts of data and identifying connections and correlations, which makes them excellent at instant trendspotting, says Daniel Anstandig, CEO of the enterprise tech platform Futuri.

  • Futuri's AI application TopicPulse takes real-time data from Facebook, Twitter, Instagram and 100,000 new sources and looks at early signals to determine what will be popular in a particular market or demographic up to 24 hours from the present.
  • A newsroom or a brand content marketer "could look at a specific audience and determine where their investment of time and energy in terms of content generation is going to be best used," says Anstandig.

Between the lines: AI can increasingly help with generating that content as well.

  • Hamburg-based neuroflash, a marketing company, uses natural language processing AI to generate scores of suggested slogans, email subject lines and social media language.
  • Its AI is then able to rank each suggestion on a 0–100 scale, based on historical data about customer preferences and calculating how different messages will resonate with different consumers.
  • Big language models "are very good at generating possibilities," says Henrik Roth, a co-founder of neuroflash. "And then humans can bring an editorial quality to pick the best one," taking into account how the AI itself ranks suggestions.

Details: I used neuroflash's product to try to generate a slogan for Axios Future, my twice-weekly newsletter.

  • After inputting a roughly 150-word description of the newsletter and selecting a few keywords — like "future," "artificial intelligence" and "journalism" — and the U.S. market I was focused on, I let the machine work.
  • The results weren't too bad. "From the future to you," ranked highest, followed by "Future-focused in a flash" — the latter being clear evidence that the AI had twigged to marketers' love of alliteration.
  • I could also shift the goals of the campaign, choosing "complex" or "simplistic" language, and selecting a specific sentiment, like "happy" or "powerful."

The other side: If this feels a little mechanical, some flesh-and-blood admen agree.

  • Rory Sutherland, a vice chairman at the advertising giant Ogilvy, compared AI recently to "satnav in a car. Great for directions but you don't allow it to drive the car."

What's next: As AI text and even video generation improves, more and more of the digital ad copy we see will be at least touched by a machine.

  • D-ID, an Israeli startup, has partnered with film studio Warner Bros. to create a promo for the new sci-fi film "Reminiscence," which stars Hugh Jackman as a PI in the future who can navigate virtually generated memories.
  • The D-ID promo asks users to upload photos, which the company's deep learning system can transform into video inserted into a scene from the film.
  • "In five to 10 years, small companies or even individuals will be able to use these tools to create short films, while startups might be able to create Hollywood-level movies," says Gil Perry, D-ID's CEO.

The bottom line: And unlike the copywriters in "Mad Men," AI marketers won't raid your liquor cabinet.

Editor's note: This post has been corrected to reflect that Futuri's AI application is TopicPulse (not Topic Calls).

Go deeper