Visitor Says – NLP Chatbot
  • 3 Minutes to read
  • Comment
  • Dark

Visitor Says – NLP Chatbot

  • Comment
  • Dark

See how to use the Visitor says node and Natural Language Processing in Tidio. This can help you trigger bot responses based on what your visitors typed when they started a conversation. 

In this article, you'll learn:

Visitor Says supports 16 languages: English, Arabic, Chinese-simplified, Chinese-traditional, French, German, Italian, Japanese, Korean, Dutch, Polish, Portuguese, Spanish, Thai, Turkish & Russian.

What is NLP?

Natural Language Processing is one of the forms of artificial intelligence that aims at recognizing and understanding the natural language used by us daily. Thanks to NLP, chatbots can reply to users' questions based on phrases typed in by users.

This allows the chat to work based on real, typed messages instead of pre-made and choosable options, making the interaction less restricted. The bot can react with a specific message, based on the words that appear in the customer's question, e.g., if the person asks, "I want to return a product," - the bot can recognize the word 'return' and trigger the appropriate reaction to it:

In a more natural example - the bot NLP recognizes the question about the discount codes:

- Do you offer any discount codes?
- Do you have any discounts running? 

and replies with a correct answer:

Without NLP, your Chatbot would have to rely solely on the pre-made decision tree flow, forcing users to choose one of the pre-defined answers.

The Visitor says node is a Trigger - something the Visitor needs to do to start a particular Automation, so this needs to be one of the first elements on the grid.

Add NLP to your bot

To add the NLP trigger- create a bot from scratch or edit an existing scenario. Then drag and drop the "Visitor says" node onto the bot's editing map and connect it to the appropriate action.

Once you add the NLP trigger - you will ask for the language you wish to use the bot.


At this moment, we support the following languages: English, Arabic, Chinese-simplified, Chinese-traditional, French, German, Italian, Japanese, Korean, Dutch, Polish, Portuguese, Spanish, Thai, Turkish & Russian

Adding the conversation language to the NLP trigger


When the language is chosen - you will be asked whether you wish to set up the NLP questions/phrases yourself or use our pre-defined datasets (templates) such as; order status. discount codes, shipping details. 


Choosing a conversation topic in the NLP bot

Please make sure to add phrases that include at least two words for the system to recognize them correctly. One word may not be enough as they are often repeatable in different bots & conversations. 

 In the last step - you'll be asked to add your datasets with the questions or phrases that users can use while writing to you through the chat. 


Feeding Your NLP Chatbot

Adding just a single phrase to a bot is not enough. With only one phrase, our Visitor would have to put in the exact words for the Chatbot to work.

One node (Trigger) Visitor Says = One Language

Remember that you can use only one language in one Visitor Says node. If you wish to trigger one Chatbot using phrases in different languages, please create separate Visitor Says triggers for other language versions or create a separate chatbot.

You can add multiple sentences to "feed" the NLP engine. Phrases added to those fields will trigger a bot response.  The more phrases we add, the greater the chance our Chatbot will recognize what the Visitor has written. We suggest using at least 3 phrases, but the more, the better!

Because we have implemented so many phrases, our Chatbot will recognize what we want even if we are not using the exact phrases:

NLP recognizing phrases


NLP Chatbots in a Conversation

The NLP chatbots will only work in a conversation not assigned to any Operators. We designed them that way to avoid any unnecessary and annoying interruptions. 

NLP in a conversation


To learn more about the basics of editing a chatbot, look at our chatbots guide

If you still have some questions for our team after reading this article, don't hesitate to contact us.

Was this article helpful?