Spring 2017 Q2 • 2022Q2 • 2022


13


assets,” says Julie Saxon, Chief Revenue Officer at IBM Watson Advertising and the Weather Company. “Regardless of how good the data sets are, marketers will still need technology to help make sense of what’s in front of them. That’s why we believe AI can be the foundational technology to advance the data narrative. But now the challenge to both open-web marketers and publishers is whether we’re willing to take the steps necessary to invest, test, and implement AI, test, widely across our operational teams. Are we keAI the industry’s f


ur operationa


ready to make AI the industry’s foundational technology, creating deeper connections with


with our audiences, or are we okay with the atus quo?”


gy, creating deeper connec our audiences, or are w ok


status quo


AI tools can create better look-alike models for prospecting, determine the most effective content marketing, efficiently set up and interpret A/B email and online tests, and help fine- tune messaging.


What Is AI? Artificial intelligence conjures up images of machines with minds of their own taking over, à la HAL, the eerily humanlike computer in 2001: A Space Odyssey. In practice, however, it’s not at all scary. In simplest terms, AI applications are


Are we ready to make AI the industry s foundational technology, creating deeper connections with our udiences, or are we okay with the status quo?


those that hav been pr


algorithms to respond to data in a way that mimics human behavior.


se that have been programmed with hms to respond to data in a way that man beha


ammed with


Let’s take a form of AI most of us are familiar with: product recommendations. A basic book recommendation program would be written so that someone who bought a book by Stephen King would, on subsequent visits, receive


am


phen King eceive


recommendations for other books by King, as well as books on similar subjects as per predetermined parameters. As this example shows, at the heart of AI are if-then and if-else statements.


s b King, as per


mple Now let’s say that, over time, a


significant percentage of buyers of Stephen King books also bought books by Danielle Steel. Given how dissimilar those two writers are, the person who wrote the recommendation program most likely did not include code along the lines of “if customer buys King, then recommend Steel.” But if the program was written to encompass machine learning (ML), a subset of AI, then it would detect buying trends such as this and subsequently “learn” to recommend Steel books to King fans and vice versa.


A subset of ML—and therefore a


sub-subset of AI—is deep learning, which requires large data sets and sophisticated programming to solve for complex


problems. Self-driving cars, Alexa, Siri, and Netflix recommendations are among the applications dependent on deep learning. Natural language processing, or NLP, is another subset of machine learning. This enables computers to interpret and analyze human language, not just code. Chatbots and voice search applications are among the more common instances of NLP in action.


a form of AI most of us are duct r ommendations. mendation pr t someone


Are we ready to make AI the industry’’s foundational technology, creating deeper connections with our audiences, or are we okay with the status quo?


AI in Use The alphabet soup of acronyms can obscure the wealth of ways artificial intelligence can help organizations market more effectively. Below are examples of five common AI applications.


Product recommendations


According to a study by e-commerce solutions provider Kibo, 52 percent of retailers already have a product- recommendations engine on their site— and for good reason. Suggesting relevant products to site visitors can not only convert visitors to buyers, but also increase average order values. Businesses have also successfully used recommendation engines to personalize abandoned-cart and post- purchase emails, boosting conversions and leading to add-on sales. Build with Ferguson, which sells home


improvement products to both consumers and trade professionals, realized that it needed to address these different audiences in different ways. Working with optimization solutions provider Dynamic Yield and its AI-powered segmentation tool and recommendation engine, the company tested multiple suggestion algorithms. It found that recommendations based on recently viewed products were most productive among trade professionals, whereas recommendations that also incorporated affinity algorithms, based on purchases made by other shoppers with similar interests, generated the best response among consumer shoppers. By better personalizing its suggestions, Build with Ferguson saw an 89 percent jump in sales of recommended products. What’s more, it found that site visitors who


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20