Reviving Marketing Automation with Gen AI: Reality or Hype?

Topping the list of Gen AI’s beneficiaries

Defined by Gartner, Gen AI is a subfield of Artificial Intelligence that can learn from existing artifacts to generate new, highly realistic content, including text, images, video, music, and even software code [2]. The technology has recently become mainstream with the launch of ChatGPT, a chatbot developed by OpenAI that made headlines for its incredibly human-like interactions. Ever since, its applications have been extensively studied by technology enthusiasts, with reports estimating that Gen AI could generate up to US$4.4 trillion in economic benefits. Specifically, marketing and sales are predicted to be among 4 functions that, together, reap as much as 75% of that value [3]. Without a doubt, Gen AI soon stands at the forefront of marketers’ strategies, with a whopping 70% of organizations already utilizing Gen AI to address marketing challenges [4].  

Personalization at scale 

Personalization has consistently topped the list of AI’s benefits and it is no exception with its subfield, Gen AI. Indeed, personalization emerges as the most popular Gen AI application in marketing, with as many as 67% of marketers currently exploring such use cases, according to a BCG survey [4]. Marketers are utilizing Gen AI for personalization efforts in several ways, with one prominent method being improving campaign effectiveness and subsequently driving customer engagement with more targeted content. What makes Gen AI highly effective in this regard is its capability to learn from millions of data points, such as past campaigns’ performance, created content, customers’ reactions, and current trends, to generate and recommend content that is tailored to different campaigns, channels, and customer segments. Unlike its human counterparts, who draft messages based on a limited data pool and analysis, Gen AI offers a significantly more data-driven approach to messaging personalization. 

Global brands have already been exploring this use case and are starting to reap positive results. For example, Michaels, North America’s largest art and craft retail chain, turned to Gen AI to generate and customize content for three channels – SMS, Facebook and email, aiming to deliver more relevant and personal content to its targeted audiences. The Gen AI tool was first fed with Michael’s existing content to learn and understand the brand voice, which then generated drafts and conducted language experiments with predictive models to learn how audiences would react to specific messages across campaigns. This analysis was then used to recommend the right message for the right channel and campaign, allowing the company to deliver more relevant content. The technology was able to predict the audience’s reactions and make practical recommendations as it was trained on an extensive pool of tagged words and phrases mapped to human emotions, benefiting from analysis of the equivalent of 600 years of consecutive A/B testing. Utilizing Gen AI, Michaels managed to personalize 95% of its email campaigns, up from 20% in 2019. The AI-generated content proves to be effective in improving customer engagement and loyalty, driving the click rate for SMS and email campaigns by 41% and 25%, respectively [5].  

Content generation with speed and quality 

Renowned for its capabilities to generate artifacts, it is no surprise that content creation remains one of the most explored use cases of Gen AI, with nearly 50% of marketers already digging deep into such an application [4]. The application has become so powerful that experts predict that up to 90% of content online might be generated by AI [6], thanks to its content creation capabilities that are not limited to text but also include images, video, music, and even software code.  What makes Gen AI have such great power in a much shorter time is the fact that its applications are widely available on the market, ready for purchase. Text generators, such as the headline-making ChatGPT, offer open access, pay-as-you-go services. The AI-powered tool has also recently updated its features by introducing GPT-4, allowing users to provide prompts and images as input to generate creative and technical writings, such as scripts, songs, and blogs. Its sibling product, Sora, also sparks enthusiasm worldwide for its capabilities to generate highly realistic videos from text inputs, meaning that users can provide simple descriptions to create videos without any filming, drawing or editing. Similarly, image generators and AI-powered editing are grabbing the world’s attention, with technology giants such as Adobe incorporating Gen AI in their services. Firefly, developed by Adobe, allows users to generate images, fill images with AI-generated content, create vector graphics, and edit styles and texture, all from simple text prompts. Even more, Gen AI can even generate software codes and even complete software products from text, with technology vendors such as OutSystems exploring this field. 

Despite being popular only recently, Gen AI has been enthusiastically experimented by global brands. For instance, Coca-Cola launched the Masterpiece advert, combining AI-generated animation with live action. The advert was co-created with OpenAI, utilizing its DALL-E2 generative image model and ChatGPT to animate several world’s famous works of art and combine with the Coca-Cola brand [7]. Similarly, Mattel, an American toy manufacturer, has also turned to Gen AI for inspiration, utilizing the technology in developing its Hot Wheels product. The company managed to generate as many as 4 times more product concepts compared to its pre-tech solution and use this content for design and feature development [3]. Conversational bot is another popular use case that strives to offer human-like interactions with users in contrast to the traditional script-fed, automated bots. Duolingo, for example, has recently leveraged GPT-4 for its chatbot, offering immersive features such as role-playing. The new feature allows users to interact with an AI-powered persona and complete learning tasks, such as ordering coffee with a Parisian AI barista. For a highly engaging experience, the characters are developed with their own back stories and characteristics, which can be uncovered through deeper interactions and conversations [8].  

Generate with cautions 

Though exciting as it gets, Gen AI is sparking controversial discussions among tech enthusiasts, as more than 1,000 tech leaders and researchers signed an open letter to call for a pause in AI development, fearing the technology to present “a profound risk to humanity and society” [9]. Unquestionably, such a fear is not baseless as AI applications in general and that of Gen AI in particular, when in the hands of wrong intentions, could be a source for criminal activities. The quality of AI-generated content has become so realistic that it seems impossible for the naked eyes to differentiate real from artificial content. To make matters worse, the general public is not even aware of deepfake technologies and their impacts, as reported by 72% of respondents in a UK survey. This means that identifying forgeries would now be even more challenging as the audience is unaware of their existence and how it works [10]

Given the urgency for AI governance, governments worldwide are starting to introduce regulations for utilizing this technology. For example, the European Union proposes that AI-generated content must follow certain minimum requirements, which may include marking the content as deepfake to ensure users stay aware when dealing with content manipulated by AI [10]. Similarly, some US cities and states have already implemented laws restricting AI use in certain areas, such as hiring [11]. As the regulation of AI is escalating fast, it is important that marketers keep an eye on the regulation changes to ensure the safe, secure, and compliant implementation of AI. 

Author Nguyen Vu Quynh Trang