Marketing Meets Machine Learning — The function and future of AI in marketing

Feb. 15, 2024

By: Krystal Wu

 Courtesy of spiceworks.com

Artificial intelligence and big data have been all the rage. You've probably come across phrases like "data-driven marketing" and "predictive analytics" countless times within and beyond Marketing Society, been told how useful a data-oriented skillset is during coffee chats, and listened to debates over the ethics of ChatGPT to the point of exhaustion. You get it — we all do.

Certainly, it is incontestable that AI stands as a progressive and groundbreaking force. A McKinsey report published in June 2023 suggests that Generative AI, if used effectively, could potentially add trillions of dollars in value to the global economy. But beyond economic speculation, job search anxiety, and general hype, what does any of this actually mean for the marketing and sales industry, and how exactly is the technology being implemented in the present?

Now you see it, now you don't

Although some forms of AI marketing technology — such as chatbots and image generators — are obviously discernable, the true fascination lies within AI's often invisible nature. Some of the most frequently used AI-driven tools include search engine advertising solutions, email marketing platforms, e-commerce solutions, and content creation tools. With that being said, these tools do not fall under the category of "general" AI — autonomous machines with the capacity to think and communicate in a manner similar to humans. 

True connection with consumers in marketing still requires a human touch — storytelling, branding, and empathetic customer experiences are ways in which this most prominently comes to light. As such, AI in marketing typically refers to tools or software that assist in carrying out specific tasks, such as personalizing an email newsletter to increase click rates. These tools embody intelligence in that they are designed to deliver increasingly better results as they are exposed to more data on consumer behavior and past performance metrics. In other words, when fed with appropriate data, they independently learn patterns and trends. Subsequently, they devise ways to optimize current strategies and predict the performance of those in the future.

Another notable detail is that AI in marketing is often implemented ad-hoc. Numerous marketing departments currently lack a coordinated strategy to integrate AI into big projects, which may speak to how a data-first culture is still lacking in the marketing industry (Forbes). Other industries with a lot of data and a high need for intelligent automation, such as healthcare and financial services, have been using AI for the last decade. Nonetheless, even in these industries, AI is primarily used within overall business operations instead of marketing and sales functions.

 Courtesy of mckinsey.com

Where AI is right now

Although AI is currently underutilized in the marketing industry, a plethora of tools and resources are being leveraged for targeted purposes. Here are a few examples of how AI is already making an impact across various marketing sectors:

Advertising

• Google and Facebook's online ad platforms provide companies with tools to combine audience segmentation — splitting customers into groups according to demographics such as gender, age, or income level — with predictive analytics, forecasting which customer groups a particular product/service is most likely to appeal to. Businesses are then able to target potential customers with different versions of advertising materials, boosting sales through digital advertisements.
  

Courtesy of wordstream.com

PR

• The PR industry centers around securing coverage for products and services in mainstream and specialist publications or directly through social media and third-party content creators. In this respect, AI can aid in curating potential influencers or creators whose audiences or skills align with a brand’s appeal and values. Other AI tools aid in achieving the same goal by writing press releases, shaping external messaging points, or researching the best outlets for gaining coverage. Prowly, for example, is a tool that generates personalized press releases after the user answers real-time questions. ChatGPT, of course, is also a favorite for generating written content for campaigns.

Courtesy of prowly.com

Content marketing

• Content marketing focuses upon ensuring a business's digital and in-person content establishes its brand, positions itself as an industry expert, and ultimately generates leads and sales. While the use of AI in content creation and metric tracking is widespread, a notable example is Buzzfeed’s successful utilization of AI. As one of the biggest content-driven sites in the world, Buzzfeed uses AI to drive every facet of its operations. Their AI tools have the ability to estimate how likely a piece is to go viral, suggest content to visitors according to their interests, and automate aspects of publication such as keyword selection, categorization, and personalization. This is a prime example of a very strategy-focused, AI-driven content outlet.

Courtesy of hubspot.com

Email marketing

• The basis of email marketing is pretty straightforward, and it is mostly built upon tweaking headings, scheduling emails, and copywriting effectively in order to positively impact open and click-through rates. Phrasee, for instance, is an AI platform that automates the creation of subject lines and copywriting. It also produces analytics for professionals to learn which words, emojis, syntax, and sentiments resonate with their audience. Another example is Seventh Sense, an AI software that optimizes the timing of email deliveries specifically through HubSpot and Marketo. Emails are automatically delivered to each recipient at their most optimal engagement time — when they are most likely to check their inbox and engage with messages.

Courtesy of phrasee.co

Does potential equate to impact?

With the abundance of ways AI is currently being implemented in marketing and sales, the true answer to that question likely depends on businesses understanding the need for a coordinated, strategic approach to implementing AI in their marketing departments. As marketing delivers its impact through a combination of numerous segments, these tools would only be truly powerful when used in tandem. Blindly investing in AI for the sake of keeping up with industry trends is a setback, not a breakthrough. Doing so, for one, may result in generating solutions that do not align with business goals, add value to business operations, or address real world problems. As marketing campaigns change, moreover, it would become significantly challenging to integrate previous AI solutions into new marketing strategies without a clear understanding of how the AI operates and adds value. In the process, a significant proportion of time, budget, and expertise would doubtlessly be drained.

Another crucial aspect for businesses is recruiting employees with skills relevant to the AI technologies they use. Overall, AI is designed to enhance and streamline human tasks and processes. Optimally, it should enable marketers to upskill and leverage technology to advance their existing tasks of analyzing data, automating repetitive tasks, personalizing marketing campaigns, and gaining insights into consumer behavior. This is arguably more important than hiring more data scientists in marketing departments or even replacing certain job functions, which would be both costly and risky for maintaining a seamless cross-functional department. Moreover, AI is notorious for generating content or insights with inaccuracies, biases, and divergent tones. Human oversight is crucial for identifying these flaws, correcting the resulting mistakes, and deciding how to enhance AI technologies to minimize error rates. 

This is not to mention all the barriers to entry and impact which AI technology brings. Some professionals still perceive AI as an abstract, sci-fi esque domain, overlooking how approachable and accessible it actually is in a business context. Others have glorified AI as a panacea of the future, or chosen to use it without learning the foundations of AI and machine learning–a fault termed "black box AI," where you receive output from AI without knowing how it was produced behind the scenes. There is also the all-pervasive issue of data privacy. As marketing and sales have become increasingly personalized, customers are also valuing their data privacy more and more. AI that requires data concerning customers' previous behaviors, access to cookies, etc. may face challenges in this respect–and businesses must ensure that their AI software complies with privacy laws.

Courtesy of hallakate.com

From an ethical approach, we could speculate upon whether AI should be used for certain marketing functions at all. Marketing can be very formulaic, but it is also loved and celebrated for creative campaigns and the people who spearhead ingenious strategies. Automating specific tasks in marketing is one thing, but developing a master algorithm or software to unlock the perfect marketing formula is another. If marketing becomes mostly automated, the concept of creativity is threatened, regardless of how positively it may impact businesses' sales revenues or KPIs. And if the use of AI is leveraged across creative domains for the sole purpose of increasing revenue and delivering larger numbers, could this truly be defined as innovation?

No matter your viewpoint in this conversation, you will play a pivotal role in deciding the answers. If you are a marketer, you are the future.

---

Krystal is a sophomore majoring in data science at CAS. Her (slightly obsessive) interests include music, fashion, media, creative writing, and poetry.