To NLP, or not to NLP, that is the question.
Chatbots are popping up in just about every business these days. They’re becoming a staple in marketing campaigns and developers are constantly experimenting with different approaches to make chatbots mainstream.
While plenty of platforms insist that you can use them to build a chatbot in ten minutes or less, the time and effort you put into it define whether you create a borderline human conversation or “just another dumb bot”.
The most common approaches in the bot development market include the simple Directed Dialogue, and the more sophisticated NLP. Here’s a quick breakdown to help you decide how to build your best bot.
What is Directed Dialogue?
Short story: You write the dialogue for your bot. Long story: You take a chatbot, define its character, the queries users would input, and then program an output for those queries.
Functional? Yes. Amazing UX? Barely.
These are the types of chatbots people easily stump (and then make fun of). Even so, they can be the humble customer support rep who can only help if you ask them what they’ve been trained for.
When to use Direct Dialogue for your chatbot
If your product or service offers limited choices in the form of buttons or a list of options, then a direct dialogue approach would get the job done. It just won’t win any user experience awards any time soon.
What is NLP?
NLP stands for Natural Language Processing. In a nutshell, it’s a form of AI where a chatbot is programmed to seem more human. It considers context and pattern recognition, then responds in a conversational format that people understand. Sometimes developers add buttons to enhance the structure of the conversation.
For example:
Chatting with the friendly Techcrunch personalized news bot.
An NLP bot tries to talk like we do, and when done right, it can do wonders for a business. Check out KIA’s NiroBot for a glimpse of what a well-designed chatbot can achieve.
When to use NLP for your chatbot
Bot developers will tell you that it depends on what you want it to accomplish. While that’s true, you can still more or less discern when to use NLP for your chatbot.
If your users have too many options to choose from (menu items, color options, etc.), or they can ask multiple questions about a focused subject; then NLP will ease their search for information. This makes NLP bots ideal for retail, banking, academic purposes, and all kinds of campaigns.
NLP technologies are certainly still in their infancy, but you can find a roundup of a few great NLP chatbot tools in this post. Enjoy.
Developing a chatbot involves a lot more than just defining the science behind it. For a more in-depth guide on them, there are dedicated bot developers willing to spread the knowledge.
You can try your luck with them in forums, or you can meet them in person at our upcoming VOICE conference here.
