The promise of natural language processing has been decades in arriving, but now it is here, it opens up a new world in digital communication. The arrival of Siri, Alexa and other talkative assistants is all well and good, but there’s no point if the device or service doesn’t understand what the person is asking for, or talking about. Over the few years of their short lives, they have become better at understanding people thanks to natural language processing (NLP), a technology that learns words and phrases and their meaning to provide a better service.
With chatbots now steaming onto Messenger, Skype and other platforms, acting as a 24/7 concierge or helper there are even more reasons to be aware of these technologies and the role NLP plays in them. Chatbots usually run on a script, but if the human on the other end isn’t following that script, then the conversation will run dry very quickly. NLP allows the chatbot to spot key phrases and go with a different approach.
The likes of Google’s Cloud Natural Language will drive all kinds of services in the cloud, as it and other services scan through documents or conversations for syntax, entities, sentiment and context. Soon the machines will understand everything we write or say.
While the focus is on the technology, at the backend, there are now large numbers of people being employed with Machine Learning skills to help develop new systems and drive forward this part of the digital economy.
The Inevitable Rise Of Digital Assistants
Between them, voice assistants can help with task automation, reminders and find your favourite music. Chatbots can help people engage with companies, services or other businesses, especially when there’s no one in the office, or when the customer doesn’t really want to speak to a person.
But we’re still in the first generation of this type of digital assistance, and progress will be gradual and harder to spot. As the tech companies drop the 2.0 and 3.1 tags that were such a part of tech evolution. Now, the assistants are getting better in 0.001 daily updates, through self-learning and AI technologies that just make them smarter without the need for firmware updates.
NLP will play a key part in these advances. Especially for chatbots where the conversation is often being created by a busy coder or marketing department who are learning the ropes as they go. Over time, chatbots will learn themselves, or from other bots, via a cloud-sourced list of question types and responses. But, for now, writing the right questions and preparing for all possible answers is still tricky.
At some point, psychologists or specialised designers will have to take over as conversations become more complex. That’s as people get used to bots and expect more of them, and as companies look to use them for more than just basic queries.
Between these designers, AI and NLP, bots and assistants will learn what customers are trying to say, and help get the right answer, or to the point more rapidly. While some of the public will live in fear of this type of technology, (thanks, Black Mirror!), for others, it will help speed up conversations with all types of service from travel and weather to banking, entertainment and shopping.
The NLP-Chatbot Grows Up
In SnatchBot, a free-to-deploy and easy-to-create bot tool for any type of business, NLP is a recent addition. If it sees a particular word or phrase, it can skip the current part of the chatbot’s script and move directly to something more appropriate. So, if the bot is still asking “how are you?” and the customer says “I need an appointment” or “where do I find batteries?” it can jump to a booking section or direct customers to the right section or web page.
Just as importantly, if someone starts using negative sentiments or emotions in their chat, the bot can be directed to a more conciliatory tone and bypass any “jolly” phrases that might not help the situation or the user’s demeanour.
Stay tuned over 2018 as new developments make our technological buddies that more useful and be prepared to adopt them at your business as people expect these kinds of services to offer up the latest in information and be a common utility.