Check out even more about personalization and artificial intelligence with FUSE Enterprise.
There’s been a lot of buzz in the marketing industry — not to mention in the mainstream media — about the growing power of “machine intelligence.” But the science of — and fascination with — artificial intelligence has been around for decades.
Computer scientists and psychologists have long expounded on theoretical concepts and explored advanced modeling techniques in the tireless pursuit of creating machines that “learn.” What has caused the field to explode in recent years is the exponential growth in computing power available to inventors, researchers and innovators in the subject area.
A legendary theory in computer science circles, known as “Moore’s Law” after renowned researcher and Intel co-founder Gordon Moore made the initial observation in 1965, predicted that computing power would double every two years. And it largely has, despite computer scientists boldly (and wrongly) predicting the “death” of Moore’s Law every now and then. This power has opened up tremendous opportunities to leverage machine learning algorithms in an incredibly broad range of use cases, not just a sensational event like a computer beating a human expert in the world’s most complex game, but also in things we experience every day.
Think about it: Machine learning is all around us. Apple’s Siri, Microsoft’s Cortana and Amazon’s Echo (“Alexa”) — voice recognition powered by machine learning. Self-driving cars? Coming soon. Facial recognition models used by the TSA? Already here. Even some of the news we read is generated based on algorithms, with little to no human editing required — sports results and financial activities, in particular, are fields where we have seen automated reporting emerge.
Marketing, of course, is no stranger to the benefits of machine learning. When Amazon recommends to you that “customers who browsed for this product also browsed for X, Y and Z,” it is the outcome of sophisticated machine learning algorithms sifting over enormous amounts of data to recognize patterns of product browsing across their immense customer base. And marketing organizations themselves in recent years have started to see a rapidly growing staffing demand for that most elusive of creatures: data scientists. These are the experts who wrangle big data, develop machine learning algorithms, and use the model results to drive action, supporting the most challenging problems facing marketing, customer service, customer experience and more.
As one industry pundit said recently: In marketing, “data science is having a moment.” Similarly, “predictive intelligence,” referring to a broad category of data science and machine learning techniques used to forecast outcomes, has been the rage in the last couple of years. Note that tools and technologies exist today such that the execution of predictive models can be partially or fully automated. For example, Adobe recently announced that they are rolling out an in-house artificial intelligence platform, Sensei. As TechCrunch reported, “Sensei uses various forms of artificial intelligence … to build models and distribute the data generated by these models across the platform where it’s needed.” This, combined with ongoing technology innovations and Moore’s Law, is starting to make the actionability of customer data at scale a reality.
In fact, there are even some who predict that the job of data scientist will disappear in the not-too-distant future, replaced by powerful algorithms. Gartner is already saying, “more than 40 percent of data science tasks will be automated by 2020.” That’s only three years from now! We’ll all be out of our jobs! It’s stirring up dystopian visions of machine intelligence taking over the world. There is no spoon.
But this is over the top. Despite the name, “data science” is not just science — it’s both a science and an art. And real, live humans are uniquely positioned to juggle both of these competing skill sets — to provide the expertise, judgment and, most importantly, creativity needed to tackle a non-trivial, real-world business problem and provide meaningful recommendations. It’s one thing for a tool to automatically send a follow-up email when an online cart is abandoned, or filter an email campaign list based on a preconfigured model score. But these are small things — what about the big, strategic challenges we regularly face in our marketing organizations?
A predictive model provides one piece of information to apply to a strategic business problem. Sure, it’s one significant piece of information based on rigorous scientific methodology and fueled by factual data about actual customers, not theoretical entities. But what’s the context?
Human intelligence and creativity is still vital to interpret what the model produces as output and consider it in the broader context of the business question. A model can’t know the people that will be affected — both customers and employees. A model can’t take into account the political implications of the decision to be made, and the steps needed to smooth the way. A model might be able to predict a customer’s need state based on their historical behavior, but what if the customer doesn’t view their current interaction in the same way as the model does? There are upper bounds to how much artificial intelligence is capable of when it comes to context, now and for the foreseeable future.
In summary, predictive modeling can be automated — but true predictive intelligence will never be. Predictive modeling — and all of machine learning — should play a role that empowers, not replaces, the person responsible for using models to take real action in the business. Despite the hype around machine intelligence, you'll get the best results if there's a real, live human being in the mix to interpret the data science outputs correctly — and to apply them in a meaningful way within the context of your organization.
Learn even more about the convergence of technology and branded content at the FUSE Enterprise summit. Artificial intelligence and personalization will be featured among many other techniques and technologies.
Jim Sawyer is Chief Scientist at Elicit. The company's resident savant, Jim is responsible for the artistic application of Elicit’s customer science. From evaluating the state of customer data and analytics systems to developing customized segmentation, Jim leads a team of data scientists to bring customers to life through data. He has over 20 years of experience in analytics, a Stanford B.A.S. in Mathematical and Computational Sciences and a Georgia Tech M.S. in Industrial and Systems Engineering. Elicit's Fortune 500 clients include Southwest Airlines, Fossil, GameStop, Sephora, BevMo!, HomeAway, Best Buy and Pier 1 Imports.