Natural Language Understanding

Take a closer look at the artificial intelligence subset that powers all chatbot conversations.

Talk to chatbot experts

Natural language understanding (NLU) is the technology that enables a chatbot to understand the messages it receives from humans. It is a subset of natural language processing (NLP) that utilises advanced artificial intelligence algorithms to understand and interpret the subtleties in language.

Talk to our experts

NLU through the history

In 1964, MIT students attempted to create the very first natural language understanding algorithm. The technology evolved rapidly over the next decade, starting with Joseph Weizenbaum's notorious ELIZA chatbot in 1965 all the way to Michael Dyer's BORIS system in 1983.

Today, variations of NLU are present in most NLP engines -- including ubisend's!

What about natural language processing?

While NLU focuses on understanding what the input means, NLP has two very different jobs. The first is to dissect what is said (e.g. is it a question? Small talk? What is it about? Weather, Manhattan, job opportunities?). The second is to process this information and return an answer that makes sense in the context it gathered.

In enterprise-grade chatbots, natural language process and natural language understanding work hand in hand to enable the program to understand human input.

Learn more about natural language processing.

Context

NLU enables a chatbot to understand the context behind a message, instead of taking each message in a vacuum.

Sophistication

NLU enablers a chatbot to understand sophisticated and intricate input from humans, instead of 'simple' phrase matching.

Unpredictable inputs

NLU enables a chatbot to deal with unpredictable inputs such as spelling mistakes, cursing, poor grammar, etc.

Talk to our experts

Who creates (and maintains) a chatbot's NLU?

Because NLU is a subset of a chatbot's natural language processing, it is typically implied that your chatbot's NLP provider also handles its NLU.

Companies like Facebook, Amazon, and Google have their own NLP (and thus NLU) engines. If you, or your chatbot development company, decide to use them, you must understand that all data gathered by your chatbot will also go to these tech giants. This is often a source of legal headaches.

At ubisend, we build bespoke NLP and NLU engines for our clients. These algorithms are trained to deal with your specific use case. Since NLU is all about understanding the context of each interaction, a bespoke approach to this part of your chatbot is extremely valuable.

On top of that, our approach means you know precisely where your precious data is going.

ubisend four dots

creating chatbot nlp engine

ubisend four dots

chatbot nlp training management suite

NLU is the key to professional chatbots

To build a professional, enterprise-grade chatbot, natural language understanding is a must-have.

NLU handles all the subtleties in language, misspelling, and, most importantly, context. Without strong NLU algorithms powering your chatbot, all you have is a fancy phrase-matching conversation.

There are many complex aspects of developing high-end chatbots (like we do). One of them is creating NLU algorithms that support your chatbot's one true goal and enrich your users' experience. Only a tailored approach to chatbot building will make this possible.

"ubisend’s fresh and enthusiastic approach to the task left us in no doubt that we had picked the ideal strategic partners to work with"

Chris Amos, project manager at Archant Limited

Lesser-known features of HR chatbot

HR • Webinars January 28, 2021

Webinar: Behind AI - the Lesser-Known Features of HR Chatbots

HR • Webinars January 28, 2021

Webinar: Company Culture and Chatbots - Growing an Atmosphere with AI

Overflowing HR shared services inbox

HR • Webinars January 28, 2021

Webinar: Shrinking your Inbox - Using Chatbots to Reduce Employee Enquiries

New to ubisend?

Popular links

About ubisend

ubisend ISO27001 certification