This blog is based on a published Chapter titled “Marketing and AI: Personalization at Scale.” Learn more about the book here.
The impact of Artificial Intelligence (AI) on businesses, consumers, and society in general is likely to be like that of the Internet – broad with new business models. The focus of this article is on how it is likely to impact marketing specifically, given a confluence of different factors. The Internet brought unforeseen changes such as the emergence of a behemoth like Amazon, a concept that was unthinkable during the early stages of Internet’s predominant use for communication. Before getting into how AI will impact marketing, it is helpful to understand what it is.
The term AI was coined in 1956 by John McCarthy, to describe how computers could emulate human capacity for decision making. It took many decades for us to get to a stage where computers are able to create decent music and solve problems. At the highest level, is the computer that beat a human in Go, a game tougher than chess. Such a feat requires not just problem-solving skills in the form of pre-fed questions and answers, but actual learning as humans are capable of. Human brain has evolved over millions of years and the human baby is capable of more learning than a computer is capable of at this time. However, AI is learning to become better. To learn, it requires a lot a data and a lot of computing power. This is where the concept of machine learning which is integral to AI, becomes relevant.
“The problems of pattern recognition, classification, and prediction that AI is geared to solve are applicable to many marketing applications such as segmentation, lead qualification, and predicting the likelihood of purchase“
Ramnarayan, S. 2021, Handbook of Research on Applied AI for International Business. IGI Global.
Machine Learning
Machine learning involves data, algorithms, and computing power. Effective machine learning requires a lot of data and well-defined algorithms. Algorithms in the simplest terms are rules to describe, classify, and then make some predictions. The quality and quantity of data, as well as the way the specific algorithms are defined are important here. One may foresee competition in the future through who has better data and better algorithms. Thus, machine learning is key to a computer becoming better at learning and solving problems. Here, companies have run into many challenges, such as how can you make sure that a computer can differentiate between a chocolate chip muffin and a chihuahua? Such challenges not withstanding, we as consumers are already using AI for different tasks such as using Siri or Hey Google or Alexa for making appointments, requesting songs to play, or to provide directions.
Why AI is a Necessity Today
We all feel the pressure from having to deal with keeping up with information form many different channels and in being responsive through many different channels. Each of these interactions, along with information overload creates a by-product of data. As we venture into an era of internet connected devices, we are looking at an amplification of volume and variety of data. If marketers are feeling overwhelmed by the variety of social media channel data today, imagine how much more data is likely to come their way as we have more connected devices. Making sense of such volume and variety of data and reacting to it in a timely fashion is humanly impossible. AI provides a solution here. The problems of pattern recognition, classification, and prediction that AI is geared to solve are applicable to many marketing applications such as segmentation, lead qualification, and predicting the likelihood of purchase.
AI’s Applications in Marketing
AI applications maybe categorized based on their capability to learn and respond, to predict, and based on whether they are doing physical (such as Roomba the robot for cleaning) or cognitive tasks such as Siri from Apple or Alexa from Amazon (although such technologies could be integrated in mechanical devices. They can also be categorized based on other human like responses of recognition, understanding of language, and apparent understanding of emotions. Each of these capabilities requires a different type of learning and capability and each of them has different applications in marketing.
Imagine a marketer who is free to pursue client relationships and prospects who are more important while an AI robot answers product inquiries through emails or responds through chat until the prospect or customer is deemed important enough to be elevated to a human. While chat technologies and email make it possible to be accessible to a consumer 24/7/365, it is not possible for a human at the other end to be responsive to the same extent. This is the gap that AI fills to make a marketer more effective and efficient.
Amazon’s “Just Walk Out “ technology, although seemingly simple on the outside, required overcoming immense challenges in data processing and the ability to solve challenges posed by human behavior within a store. This technology makes it possible for a store to be open 24/7/365 with round the clock access to consumers. This technology is based on computer vision that captures video images and processes them to make it possible for the customer to automatically check-out without a cashier.
On the mechanical front, both Amazon and other retailers are using robots that can pick and pack for online order and store pickup, made more popular by the COVID 19 crisis. It turns out that a robot can pick and pack in 15 minutes as opposed to 60 minutes required by humans. It can also pick 800 items per hour as opposed to 80 items per hour by a human.
On the linguistic front, a simple example is that of automatic translation of web page content in different languages. One such study showed that such translation on eBay increased revenues in trade between U.S. and Latin America by 13.1%.
At the highest level of AI capability is its power to predict – predict who is likely to be greater credit risk, who is likely to be a strong prospect, and who is likely to get a stroke – as exemplified by Apple’s work with medical researchers to show how Apple Watch could make such prediction possible.
What this limited number of examples represents is a scenario in which customers are not just numbers but are defined and explained by algorithms that enable value-add, if done right. This value-add is possible 24/7/365 and the capacity of AI makes it possible to do this while treating each customer as an individual understood based on millions of touch-points, thus enabling personalization at scale. The ability to personalize the right message and the right product at the right price and at the right time to each customer is the promise of AI. Such large scale personalization is made possible by huge volume and variety of data created by customers and individuals. The ability to harness such an explosion of data is what makes AI better, while the inability to do so is what is going to weigh down marketers. Thus, this is a partnership that is required for a marketer looking forward.
What to Expect Moving Forward
Any marketer who does not want to drown under the weight of data, should instead use that as a resource for AI. As seen in the examples above, there are different categories and levels of AI. Each of them contributes to the success of a product and customer experience in different ways. However, there are some things that are applicable to all marketers. AI makes it possible to provide value to customers round the clock 24/7/365 and to do so with personalization at scale. Marketers should be looking at how to automate anything that is a repeated action, while looking for how value can be added by creating a 24/7/365 personalized customer experience.
This blog is based on a published Chapter titled “Marketing and AI: Personalization at Scale.” Learn more about the book here.