​Artificial Intelligence: What’s Real and What’s Not in 2017

(Editor’s Note, this post originally appeared in the Silicon Valley Business Journal, and was syndicated in 42 Biz Journal publications across the country)

I’m a big Star Wars fan, so when “Rogue One: A Star Wars Story” descended on theaters late last year, I braved the crowds to see it—twice in the first 18 hours. And just like all the other Star Wars movies, “Rogue One” stoked our geeky imaginations with all the technological possibilities of a galaxy far, far away, like holographic displays and all sorts of strange devices.

And did you notice the Imperial server farm? Of course, advanced artificial intelligence (AI) was well-represented too: Like C-3P0, R2-D2 and BB-8 in earlier movies, Rogue One’s K-2SO displayed uncanny humanness.


The futuristic Star Wars-esque world is still mostly the stuff of Hollywood movies, but technology visionaries are hard at work bringing us ever closer. AI, or “ intelligence exhibited by machines,” is one area that is evolving into reality, and there are some subcategories under AI with practical applications that we use today. Natural language processing (NLP) and machine learning are two of them, and their potential for the future is exciting, especially for B2B technologies like customer relationship management (CRM).

The ultimate destination of AI for the business world is to make people’s jobs easier. Just like K-2SO in Rogue One, AI will become our intelligent personal assistant that saves the day and makes life easier. However, you may have noticed that no one has a droid at the office yet. So, let’s look at what technologies are real right now and what is coming farther down the road.

What’s real: Natural language processing and machine learning

NLP technology is quite adept at decomposing language into parts, understanding the baseline intent of that language, and representing it in both spoken and written word. Perhaps the best-known application of NLP today is Apple’s “Siri.” Ask Siri any number of things—“What time is it in Berlin?” or “When is my next meeting?”—and it can tell you.

NLP has also found a foothold in CRM. Intelligent CRM tools strive to improve customer relationships, and use NLP to mine for topics of interest from customer conversations in email and CRM. What better way to connect than to personalize customer communications? Many agree, and subsequently, NLP has benefited from a lot of investment lately.

Another subcategory of AI that is much more real than ever before is machine learning. Its recent success is due to the availability of huge volumes of discrete data points, and with this deluge of data at the ready—including digital data like social data, data from public records, and IOT data—patterns begin to emerge once analyzed.

Machine learning uses algorithms to understand patterns in data sets, and then applies some logic to the patterns. (“If ‘A’ looks like ‘B,’ and ‘B’ looks like ‘C,’ then ‘A’ also looks like ‘C.’”) Machine learning algorithms are also self-learning and have been designed to take feedback, which means that their “intelligence” grows as they analyze more patterns.

In the last few years, machine learning has made a particularly big splash in image and video recognition. Some visual recognition algorithms can analyze pictures from the Internet and understand the emotional intent behind the picture. This capability is especially valuable in commerce when a brand wants a greater understanding of its levels of customer satisfaction.

For example, one machine-learning technology trawls different social networks, looks for its customers’ brands in photos, and discerns the mood of the people in the picture with the brands. Is the person holding a can of Coke in the picture smiling or frowning? To be clear, the algorithms don’t understand the emotions of sad or happy, but they understand the difference between a mouth that is turned down versus one that is turned up in a smile.

While this latest advance is certainly impressive, machine learning’s greatest strides are yet to be made, and CRM in particular stands to benefit significantly.

What’s not real… yet

In 2013, the world watched as IBM Watson leveled its human opponents on Jeopardy. The idea that a computer powered by NLP and machine learning algorithms could spit out correct answers to so many varied questions was a curiosity. Be fair, though: Watson had a huge data set—the Library of Congress—at its—er—fingertips.

Keep in mind that machine learning’s success practically applied has only come about relatively recently, thanks to the advent and use of SaaS and cloud platforms, which can cost-effectively collect massive amounts of data. It’s also taken awhile to collect, publish and aggregate enough discrete data points in which to find and analyze meaningful patterns.

Organizations in many industries have already found a way to use machine learning to their advantage. In the coming years, CRM, too, will be primed to truly take full advantage of machine learning. Companies have by now collected trillions of rows of customer data to find patterns, and are beginning to train algorithms around how customers act. These algorithms are learning about what customers are most likely to do next based on their behavior patterns.

Tomorrow’s CRM system will be more than just a database. It will capitalize on machine learning to become the ultimate personal assistant. It will not only make a user more efficient and effective at getting the job done, but will also reveal something the user didn’t already know about his customers. This is where machine learning comes in, mining massive amounts of social and other data to uncover unknown details about a customer that will deepen the customer relationship. Tomorrow’s CRM system will also apply AI to supplement declarative rules based on workflow systems that will make predictions as to what a user should be doing next.

Back to Watson. IBM Watson provides a model for CRM to even go beyond the “ultimate personal assistant” with its personality profiler service. The service needs only an email address to scan all content that can be attributed to the person behind the address—every blog article, every tweet, every Facebook post—and then determine the personality characteristics of that person. What would that capability mean for CRM—for the sales engagement process?

The type of personality-rich and amusingly expressive cognitive intelligence displayed by K-2SO is a long way off. But the AI revolution promises new and exciting use cases in the very near future, and holds great possibilities for CRM in particular. With this in mind, in 2017 and beyond, organizations must begin to view their CRM systems as much more than just a database; rather, it must be seen an intelligent tool that has the potential to transform customer relationships.

Subscribe to SugarCRM Updates

Sugar’s email newsletter is filled with the latest trends, tips, best practices and company news you need to help drive extraordinary customer experiences.

I have read the SugarCRM Privacy Policy and consent to the processing of my personal data.