It’s clear that moral panic over new technologies is a boon to journalism. Scary headlines about menacing robots and data breaches drive clicks, which in turn sell online ad impressions. Pre-internet, in the print world, those same headlines boosted print readership.

That was the case almost three decades ago, when respected scholar and author Nicholson Baker famously had a meltdown about the digitization of our nation’s library card catalogs. What Baker feared was a loss of contextualization. What he forgot to consider is that the future is hard to predict. In this case, he underestimated the advent of Big Data and machine learning to create book content contextualization. It’s called artificial intelligence (AI).

Today there is a similar panic about AI itself. Some applications are legitimately alarming. Deep fakes have the potential to go viral and destroy political careers, and possibly entire democracies. Autonomous vehicles, while safer than human drivers on average, still can go spectacularly wrong with lethal consequences.

There is one set of AI use cases that has mostly flown below the radar of breathless headlines. It’s the role AI is playing in the present and future of digital marketing. Consider the following – all ways that AI is making our jobs easier and our efforts more effective.

Predicting User Behavior

Similar to the contextualization of book content that Nicholson Baker feared would evaporate with paper index cards, rich data about customers, including everything from social media engagement to past purchase behaviors, have helped AI produce extraordinarily predictive propensity models. These facilitate things like suggesting various products to the customers and improving their services. The outcome is greater customer satisfaction and increased profitability.

Personalization of User Experience (UX)

Personalization can be bucketed into two categories: marketing personalization strategies (digital marketing campaigns in the form of email marketing, SMS texts and in-app notifications, all customized based on data) and experience personalization strategies (web and mobile optimization based on data). Where Big Data exists, you can bet that AI is there to help. And it certainly has in these instances!

Both types of digital marketing strategies rely on the segmentation of users – identifying target audiences. AI can create better segments by testing every combination of user traits and finding those that, in a combination described by an algorithm, deliver the best outcomes with every message and changing experience.

What sort of outcomes? Whatever you tell the AI “machine” is important to your business. It could be as forward-looking as lifetime value or as immediate as real-time predictions of who will buy this week’s “must sell” product for the smallest discount or click on the latest digital advertising. Another valuable real-time application of AI is personalized product recommendations, in both marketing and experience personalization strategies.

One famous AI application takes as its inputs the 50+ customer demographic and behavioral attributes, and the one or two most desired outcomes. In the past, arriving at the optimum algorithm to cluster consumers into segments would have taken weeks of repetitive number-crunching. This application arrives at the best model in minutes.

Measuring the Value of Artificial Intelligence (AI) in Improving Search and Chatbot Outcomes

The best way to measure the success of AI in internal search or chatbots is to tune a standard user behavior measurement system to the metrics that matter and conduct carefully constructed experiments to assess ROI. Download our whitepaper and know how to measure the value of artificial intelligence (AI) in improving search and chatbot outcomes.

Download Now

Search Engine and AI Chatbot Optimization

A powerful subset of improved experience personalization is AI-enhanced internal search and chatbot experiences (sometimes called online virtual assistants). TA Digital principals Jeff Larche and Josip Lazarevski explain in their peer-reviewed paper, Measuring the value of artificial intelligence in improving search and chatbot outcomes, why these experiences are perfect for implementing AI optimization:

Everyone visits a website or voice-assisted app with an agenda. Typically, digital marketers are left to guess what those agendas are by observing click paths. But when visitors interact with an internal search or a chatbot, they remove this uncertainty: They literally tell you what they’re looking to accomplish. Text-rich disclosures of purpose — “I need to find X” or “How can I do Y?” — are ideal for AI optimization … Free of the cognitive biases that come when humans moderate these encounters, AI robots are measurably better than alternatives.

The secret to this optimization is natural language processing (NLP). Progressively over time, machine-learning algorithms are trained to understand how user behavior and traits provide uncannily accurate inferences on user intent.

A Moore’s Law for Marketing AI?

Until recently, a trend from the early 1970s showed computer power doubling every 18 months. Called Moore’s Law, it observed how ever-advancing automation techniques increased the number of transistors on each computer chip at a logarithmic scale. The laws of physics have recently slowed this trend, but a new “law” may arise from improved machine learning methods.

As Nicholson Baker and legions of others can attest, the future, maddeningly, defies prediction. But evidence of the doubling of AI power in surprisingly brief intervals can be found in a truism from recent AI/machine learning data science graduates: The AI cutting-edge methods they learned at the beginning of their education have nearly all been replaced by graduation day!

This growing power in the coming years is for wise marketers to harness.