Skip to main content

Setup FAQ AI Bot

You can create an AI bot in Chakra to help your customers with recommendations of products from your WhatsApp shop. This tutorial assumes that you have already completed the following steps

  • Create a Meta Commerce Catalog and linked it to your WhatsApp Account
  • Connected your catalog to Chakra and synced the products

Create a Product Content Source​

Before configuring an AI shopping bot - you need to first configure an AI Content Source. This content source will be the source of product recommendations

  • Go to AI Content Source list Admin > AI Settings > AI Content Source
  • Click New Content Source button on the top right corner
  • This will show you a new modal
  • Lets fill the Name field with a relevant name like 'Products'
  • From Content Source Type select Products
  • This will show a FAQs section
  • Under Which Products please select All
  • Click Create button once done

This will index all your products and create a new content source around it

Create an AI Shopping chatbot​

Now lets create a new chatbot to use this content source. If you want to understand the basics of creating a chatbot, you can refer to the Add a chatbot page.

  • Go to the Chatbot list Admin > Chat Settings > Chatbots
  • Click Create Chatbot button on the top right corner
  • Rename this chatbot at the top to 'Shopping Bot'
  • Edit the Welcome node and add the following message 'Hi this is Shopping bot'
  • From the left sidebar drag the "AI Product Search" node and drop it on the panel somewhere to the right of the welcome step
  • Link the Welcome node to the "AI Product Search" node by clicking and dragging from the right connection handle of the welcome step to the left connection handle of the "AI Product Search" node
  • Edit the "AI Product Search" node to set its configuration
  • Under "Select specific content sources" select the Products AI content source we had created
  • Under stop keywords you can configure any custom stop keyword for this step if needed. All keywords are matched in a case insensitive way
  • Configure the header text with a relevant title like Product Recommendations
  • In the body text you can add a message to be sent along with the recommendations like Here are our recommnedations
  • You can optionally add a footer text with something like Type stop to end this chatbot
  • Under catalog Id, add your Commerce Catalog Id
  • Products section title can be something like 'All Recommndations'
  • You can optionally configure a 'No answer message' - This message will be sent to the user if the AI is not able to reliably find an answer to the user's query
  • You can add a thank you message step after the AI reply step
  • Save the chatbot
  • Mark the chatbot as active

Once this chatbot is triggered, the bot will keep answering user's queries until the user enters a stop keyword.

Add a chat trigger​

Now we need to configure a chat trigger that will initiate this chatbot

  • Go to Chat triggers list Admin > Chat Settings > Chat Triggers
  • Click Create Chat Trigger button from the top right corner
  • In the new chat triggers page, add a keyword 'ShopBot'
  • Under matching strategy select 'Starts With', this will ensure that when users start the chat with a message like ShopBot show me some plates and mugs please then the chatbot is immediately able to answer the question
  • Under actions, configure Trigger Chatbot action and select our newly created chatbot
  • Save this trigger by clicking Save button

Thats it, you can now test your chatbot.