How AI is Making Sentiment Analysis Easy
by Daniel Newman | November 27, 2019
Listen to this article now

Customer feedback is great. But how do you turn that feedback into meaningful customer insights? In the past, companies used things like surveys to try to narrow down a general good/bad/neutral response to their recent marketing campaign or product. Still, there is so much more information in the form of unstructured data that could help companies better understand their customers. Whether they are using social media, blogs, forums, reviews, or online news commenting, customers are sharing their opinions in tons of different ways every single day. The only issue: many of these opinions are shared in nuanced ways that traditional AI hasn’t been able to navigate. All of that is changing, however, with increasingly effective sentiment analysis.

The Basics

What is sentiment analysis? It’s a far more complex way of analyzing how consumers feel about our products and services, using not just simple words but longer sentence fragments. Yes, AI is becoming smart enough to understand the tone of a statement, rather than simply understanding whether certain words within a group of text have a positive or negative connotation. This is incredibly impactful for companies seeking to optimize their message, improve customer engagement, or even identify top influencers in their customer base.

The Opportunities with Sentiment Analysis

The possibilities of sentiment analysis are incredibly far-reaching. The types of information that AI can gather from both unstructured data and affective computing in sentiment analysis are huge. They could help with predictive analytics, predicting buyer response in the stock market, managing employee engagement, etc. In the past, surveys may have offered a “comment” section at the end of the survey where people could leave verbatim comments—and perhaps, every once in awhile, someone might read those statements and even do something about them. That is so far from the case as we head into 2020. Sentiment analysis is capable of 90 percent accuracy. That’s not a technology in early stages—that’s a technology in a state of maturity, ready to go to empower companies, employees, and customers all at once.

Reviewing text-based feedback like social media posts that have already been made is just one use of sentiment analysis, however. Technologies like Cognito actually review customer service calls in real-time, detecting human signals and offering behavioral guidance to improve the quality of the interaction. Their tools have been shown to improve customer satisfaction by nearly 30 percent, decrease call handle time by 15 percent and increase customer feedback by an unfathomable 90 percent. Yes, it seems empathy and EQ in customer engagement are hugely valuable, and AI-powered tools like Cognito—not humans!—are making that happen.

Indeed, one of the great ironies of sentiment analysis is that by removing the human from the service side of the conversation, you might actually improve the level of “care” one receives on their call or chat session. For instance, imagine you’re a customer service agent. You’ve just fielded your 10th call with an irate customer. You’re so frustrated you’re about to lose it yourself. Your EQ level is a negative 10. You have zero empathy for the person on the other end of the line; you’re focusing merely on the fact that you’re earning minimum wage to get verbally beaten by an unhappy customer on the other side of the line. AI to the rescue. Because AI has no emotions, it can never lose its empathy—it can only grow it. What’s more, it can be used in any industry or field, from banking to HR.

More Than Just Improving Customer Service

Still, improving customer service isn’t the only thing sentiment analysis can do—far from it. When combined with things like cognitive recognition and affective computing, it has the potential to save lives. Affective computing takes sentiment analysis from text to audio and video. With the right AI, car companies could use affective computing to determine if a driver is too intoxicated or tired to drive. For instance, are their eyes closed more than normal? Is their speech slurred? Are their words making sense? On a similar token, gun manufacturers could add an element of affective computing into gun safety locks, ensuring that someone is not overly angry, depressed, or anxious when picking up a firearm. Imagine what this could do to keep our streets and communities safe.

One car company, Kia, is already using basic forms of cognitive recognition in facial expressions to help make their driving experience more “fun” for drivers—turning up the radio and electing faster-paced songs when passengers seem to be having a good time. Is this technology necessary? No. But it helps the customer feel like the car “knows” them—understands them—is dedicated to providing a memorable experience.

The Future

Sentiment analysis is already happening, all over the world, using advanced AI techniques, and its full potential remains to be seen. Not surprisingly, the field of sentiment analysis as a service is skyrocketing, with numerous service options for those companies interested in employing these techniques. Not every company in every industry will benefit from the investment right now. However, I’m banking on the fact that sentiment analysis will likely be a key player in successful customer experience and engagement moving forward. Remember: As with any new technology, the value is not in the information you mine, it’s in what you do with it. The power of AI isn’t in replacing our need to understand our customers, it’s in using tools to understand them better and then act on those understandings, for the better.

The original version of this article was first published on Forbes.

Daniel Newman