What ChatGPT Means for Marketers

Since rolling out in November 2022, ChatGPT has stirred up a wide range of emotions among marketers, from excitement to fear to skepticism. The large language model chatbot, developed by the artificial intelligence (AI) company OpenAI, leverages natural language processing (NLP) to analyze and respond to queries with remarkably human-sounding answers. The still-nascent technology could have a dramatic effect on marketing and advertising.

"Just imagine the ability to analyze vast data assets in real time and recommend immediate actions, design visuals, build presentations — the list goes on," says Raja Rajamannar, chief marketing and communications officer and president of the healthcare business at ANA member Mastercard. "We're already seeing AI-driven media plans on a trial basis, and the recommendations are relatively decent. Once these technologies advance further, the potential will be incredible. We'll be able to cut through laborious tasks and focus on what matters most."

Even brand managers who are not in testing mode need to understand the potential implications on both the workplace and buyer behavior. Indeed, ChatGPT is growing faster than TikTok, amassing more than 100 million monthly users in less than two months, according to UBS research, per CBS News.

In a ResumeBuilder.com survey of 1,000 U.S. business leaders conducted in February, 49 percent of respondents said their companies currently use ChatGPT and 30 percent plan to. Nearly half of the respondents (48 percent) using ChatGPT say it has replaced workers, and 93 percent of current users say they plan to expand their use of ChatGPT.

The tech giants, of course, are scrambling for share. Meta, Google, and Snap continue to explore the most effective ways to leverage generative AI tools in their advertising programs, while Microsoft, an OpenAI investor, is incorporating ChatGPT into its Azure OpenAI Service as well as its Bing search engine. Google has clapped back with the launch of Bard, a conversational AI service powered by LaMDA (Language Model for Dialogue Applications).

So, generative AI is here to stay and will likely become even more pronounced when it comes to online marketing and communications. But does that mean brands need to use it?

AI-Powered Content Creation

A number of brands have already used ChatGPT to help create ad campaigns, including a spot for Mint Mobile starring Ryan Reynolds, the owner of the mobile phone company, who proclaims the results of using ChatGPT are "mildly terrifying." But it's only the beginning.

By 2025, 30 percent of outbound marketing messages from large organizations will be synthetically generated, up from less than 2 percent in 2022, according to Gartner.

"AI tools will increase the speed and variety of digital marketing content creation, eventually overtaking humans as the origin of most content/copy," says Adam Segall, an analyst at GP Bullhound, a technology advisory and investment firm.

As an example, he points to Jasper, an AI copywriting tool based on the same language model as ChatGPT, Generative Pre-trained Transformer 3 (GPT-3). But humans will still be in the mix. "AI needs prompts, and models need to be trained on something," Segall says.

AI-human partnerships will reduce the time spent and cost of creating content at scale, Segall adds, enabling ad agencies deploying AI to charge less for their services.

Julio Arrieta, global CMO at financial technology company Leap Financial, uses generative AI to create digital and social media campaigns in one-tenth the time and cost of previous campaigns, and reports an ROI four times that of industry benchmarks. "Whether you call it AI prompt engineering or AI whispering, the key to achieving these kinds of results is providing the right set of instructions, just as you would when assigning a task to a person," he says.

To create presentations and web content drafts in a flash Arrieta also uses Tome, an AI tool that incorporates DALL·E 2, the GPT-3 trained AI system from OpenAI that creates realistic images and art from a text description. He says that DALL·E 2 is particularly useful for sourcing ethnically diverse images, which can be hard to find.

Thinking Outside the Box

OpenAI API, which allows users to build apps based on GPT-3, is shaping up to have a significant effect on ad creative. Ryan Steelberg, president and CEO of tech company Veritone, says that marketers can tap generative AI's creativity capabilities by using out-of-the-box prompts to inspire ideas.

"What's interesting about these models is they don't have overly restrictive constraints like skills and actions that we are used to with virtual assistants," he says. "When you do a crazy prompt, it is not saying, 'Well, that is irrational.'"

Brands testing ChatGPT should set clear guidelines, train their teams, and monitor the results, says Chris Duffey, head of strategic development, creative cloud design at ANA member Adobe and author of the recently published book, Decoding the Metaverse. "Regularly review and assess the effectiveness of ChatGPT in the creative process," he says. "And adjust as needed to ensure it is being used in a way that enhances rather than replaces human creativity."

Sarah Berg, VP of marketing at the insurance technology company Matic, worries that AI shortcuts could contribute to a "loss of creative thinking" among marketers. Her team is vetting ChatGPT for content ideation, subject line A/B testing, and revision of stale copy, but to her mind there is no substitution for the human touch. "The majority of our team feels ChatGPT is something we can consult for ideation and inspiration, but any output of ChatGPT needs to be revised in our voice and tone," she says.

Segall agrees that relying too heavily on AI-generated content is risky. "For brands, flawed training data, bad actors, and ease of access/use can be severely damaging. Low quality content can dilute a brand. Misinformation can damage a brand's reputation," he says. "As a result, in-house use of AI tech should be carefully vetted and regulated."

Bespoke Approach

OpenAI recently announced ChatGPT Plus, a monthly subscription plan that provides access during peak times, faster response time, and early access to new features.

Steelberg says that as the technology becomes commercialized marketers are going to be able to add their own data to the system. "It will be just like you're building your own conversational AI chatbot," he says. "You will get all the power of ChatGPT and securely add everything about your company. This will be a definitive game changer for every business, especially marketing."

But even these versions are unlikely to replace humans in the near future. "I don't see AI supplanting humans anytime soon," Rajamannar says. "Firstly, we are at least a few decades away from artificial general intelligence (AGI). AGI engines would have things like intuition and judgment, abilities we view as innately human. Second, marketing is all about human connection."

Nevertheless, AI's role of making things easier for marketers may have just gotten ironically more complicated with the advent of ChatGPT. "As new demographics, audiences, topics, products, formats, etc., crop up, humans will need to train AI to produce relevant and effective marketing collateral," Segall says. "Consequently, brands will still rely on marketing and advertising partners to understand the mentality of their customers and what would resonate with them most."