When you think of the term ‘artificial intelligence (AI)’, what comes to mind? The Terminator or Hal 9000 in Stanley Kubrik’s "2001: A Space Odyssey?" Marketers might be surprised to know that while visions of cyborgs are probably decades away, artificial intelligence is changing the way we sell to, market and service customers — right now.

The field of artificial intelligence is often dismissed as the purview of pseudo–philosophers, science fiction writers, or hack scientists with too much time on their hands.

However, artificial intelligence applications and technologies are not confined to imaginary worlds. In fact, AI is helping companies across the globe create operational efficiencies, lowering costs and improving the customer experience.

But before we get too carried away with thoughts of machines running amok and apocalyptic scenarios, it is helpful to separate hype from reality by delineating two different types of artificial intelligence.

Strong AI vs. Narrow AI

Through movies, video games and science fiction books, most people are familiar with the concept of strong AI or machines that may someday have the ability to plan, learn, perceive, feel and intuit. And despite the strong push from many scientists and trans–humanists, ‘strong AI’ or ‘artificial general intelligence (AGI)’ is currently more fiction than fact.

In contrast, narrow or applied artificial intelligence is available today. It helps companies perform complex calculations and computations based on large data sets that would typically take humans too long to perform.

Narrow AI is at work in the business processes of many corporations. Ray Kurzweil, a futurist in the artificial intelligence community, often cites how narrow AI applications and computational machines provide assisted decision support by coordinating shipping schedules, flying and landing airplanes, making better credit decisions and recommending medications for patients based on potential drug interactions. These are, of course, just a few examples.

Advances in assisted and automated decision making are made possible by exponential increases of processing power in hardware, improved software algorithms and optimized business processes. By using mathematics and advanced statistical techniques, these systems often work in concert with knowledgeable workers, helping them to make better and smarter decisions.

Marketers, however, often don’t realize how prevalent narrow artificial intelligence is (or should be) in our daily processes.

Take for example, customer attrition modeling. Marketers know they can impact and retain cash flows by keeping customers satisfied with quality products and services so they don’t defect to competitors.

One wireless company has created a churn, or defect propensity model, to predict which customers will drop service when their contracts expire.

You might think such a model would be based on logical attributes such as number of calls to customer service, billing disputes, dropped calls etc — and it is. However, you might also be surprised to discover this carrier has refined their labyrinthine algorithm to incorporate over 400 dimensions for predicting customer churn.

While not perfect, the model is fairly accurate and is helping marketers at this company create more effective marketing campaigns to disaffected customers — reaching out to them with specialized and personalized offers — long before they leave for the competition.

Using narrow AI to address customer issues

The smart marketer knows the days of 'spray and pray' — sending out thousands of non-targeted communications to customers — are effectively over. To achieve higher response rates, marketers must develop more targeted, relevant, and timely marketing programs.

There is also an opportunity, using narrow AI, to address customer issues before they snowball. Let’s examine a scenario where a wireless customer is experiencing service disruptions.

This particular customer is a subscriber to a U.S. wireless carrier. Fortunately for the customer, they signed a service contract with a wireless company that recognizes the value of competing with narrow AI.

This carrier makes a strong effort to collect data from various customer touch points, and integrates the data into a single repository or enterprise data warehouse. This data integration allows the wireless company to have a ‘single view’ or history of all transactions and interactions across a myriad of channels (call center, Web, retail outlets, etc.).

After capturing data from disparate sources, they analyze the information and detect events – in this case, service disruptions — which then trigger an appropriate response.

In this example, the customer (let’s call him Donald Smith) has dialed the company call center with five complaints over two months about dropped calls. Each time he calls to complain, Mr. Smith receives a credit for the dropped calls. Unfortunately the dropped calls continue to occur.

The wireless carrier, however, is taking advantage of sophisticated data mining algorithms to sift through the centralized data warehouse, combing for irregularities, complaints or other anomalies. Detecting the frequent complaints from Donald Smith, the analytical application triggers a service agent to call Mr. Smith and diagnose the service issues.

At this point, the service agent reviews Mr. Smith’s service history and discovers that he is using an older handset that is not as compatible with the wireless network as newer mobile phones.

Meanwhile (behind the scenes) narrow AI applications are at work to determine the best offer for Mr. Smith. Mr. Smith’s profile is pulled within sub-seconds, and a sophisticated algorithm 'qualifies' Mr. Smith from more than one hundred potential offers. The algorithm optimizes the situation by narrowing down the best offer to five.

Of course, the call center agent cannot and should not pitch all five offers to Mr. Smith, so the algorithm then ranks the offers based on the likelihood of acceptance by the customer. All this happens in sub-seconds.

The analytical application presents the best offer to the call center agent through a 'screen pop.' The first offer of a new phone with an additional one year contract is turned down by Mr. Smith. The application takes note of this decline and learns from it, so that next time it can present a better offer to Mr. Smith.

The next offer is a winner — a lower priced phone, but no additional service contract. Mr. Smith is delighted with the new phone offer and accepts the offer verbally. The call center agent then creates an order to ship a new phone via overnight delivery, simultaneously sending Mr. Smith a confirmation email.

In this situation, everyone wins. Mr. Smith receives a new phone that works better with the wireless network, and promptly tells two friends about his positive experience. And since it costs almost five times as many marketing dollars to acquire a new customer as retain one, the wireless carrier wins too.

This, then, is the future of marketing. It is an elegant and precise symphony of computational machines and humans working together to respond (in most cases proactively) to customer events, and create customer communication strategies that are personalized, relevant and delivered at precisely the right time for maximum impact.

The global economy is volatile, technology isn’t standing still and customer expectations are increasing daily. By championing and adopting sophisticated analytical tools, technologies and applications (narrow AI), marketers will position their companies to increase their competitive edge and to succeed now and into the future.