3 reasons natural-language tech has more potential than you think

With the announcement of funding for AI-driven DigitalGenius last month, artificial intelligence, and particularly Natural Language Processing (NLP), have been making headlines recently. Customer service powered by NLP? Computers conversing with humans and actually guiding them through the sales funnel? It all sounds a little futuristic and scary, right? But the fact is, many of us are already using NLP every day — both Siri and Google Now are NLP-driven.

As exciting as DigitalGenius is, however, it only hints at the extent of NLP’s capabilities. Before we get into the future of NLP, though, let’s clear up a few common misconceptions. There are a lot of people who think they know about NLP — and even more who don’t.

Misconception No. 1: NLP is voice recognition. This is a significant misconception. Sure, Google Now, Siri, and Amazon Echo all use voice, but that’s only one of NLP’s applications. Furthermore, it’s not the best or most accurate use of the technology: NLP is much better at analyzing text-based data.

Misconception No. 2: NLP is most useful to passively analyze large amounts of data, such as when it’s used to gauge the sentiment of aggregated social media posts. Again, not true. In fact, NLP can be used to analyze and “understand” real-time data in small chunks. The next wave of NLP will likely be integrated into an application’s user interface, and that makes the future of technology pretty exciting.

Misconception No. 3: NLP is not mature enough to be really useful. It’s a digital parlor trick. This is completely off the mark. As the DigitalGenius launch proves, businesses and consumers alike can benefit from the use of NLP.

But while NLP is a driving force in the future of web and application development, its intelligence isn’t boundless, and many of the current use-cases expose its limitations. Apps that provide general functionality and intend to employ NLP like a dictionary to understand every word, intent, and idiom will disappoint and frustrate users.

The truth is that there haven’t been any commercial uses of NLP that are truly intelligent. Nothing has capitalized on NLP’s full potential. Instead, developers rely on it primarily for translation from voice into keywords, they then use those keywords to seek information from the Internet. But NLP cannot be all things to all people; it works best when restricted to a single context. Think of it like the Apple Store Genius Bar: A genius can tell you everything about your Apple product and repair it. But you wouldn’t go to the Genius Bar and ask them for the best place to get chicken wings or what to buy your sister-in-law for her baby shower. If you did, the response would probably be similar to what you’d get from Siri.

There are, on the other hand, ways to use NLP that can yield promising and even exciting results. Narrowing the scope to a single context, like customer support for a particular product, NLP can be used, along with machine learning, to not only understand what you said, but also learn from experiences what you didn’t say and what your actual intent was. An NLP-based travel app, for example, could learn your favorite airlines, hotels, and flight times through your interactions and history, and on request, make travel decisions and present you with reasonable itineraries for your next trip.

The future of NLP is bright and will lead to truly intelligent applications in the not-too-distant future. From CRM to virtual travel agents, these apps will prove far more useful than asking Siri about the laws of robotics.

Ben Cheung, CEO and cofounder of meeting-scheduling company Genee.