Updates in Chat | Yext Hitchhikers Platform

What You’ll Learn

In this section, you will learn:

  • An overview of the key features of the newly released Chat product

Chat is the newest available product in Yext’s composable digital experience platform. Yext Chat leverages Large Language Models to offer a highly-configurable conversational interface that can:

  • Automatically answer users’ questions based on your content
  • Collect user inputs (such as contact information)
  • Perform specific tasks by integrating with other systems (such as initiating a return)

Goals and Instructions

Every Chat experience is configured using Goals & Instructions. Goals & Instructions allow you to customize the behavior of your chatbot for your use case, and allow you to tell your bot how to handle specific scenarios that may come up in conversations with users.

Goals represent a general objective of the user, like “Answer a Question”, “Order Delivery”, or “Schedule an Appointment”. You can set up any kind of goal by simply providing a description of the goal, along with a few example queries that may indicate a particular goal.

Instructions offer step-by-step guidance for the bot so that it knows exactly how to achieve each goal. These steps are also configured using natural language.

For example, if you link a search experience to your bot, it will automatically set up an Answer Questions goal with these instructions:

  1. Search Yext Content for information relevant to the user’s question.
  2. Answer the user’s question based on the search results.

Yext Search Integration

The ability to link a search experience with your chatbot is particularly helpful because bots are most commonly used for answering questions based on a business’s knowledge base. For example, a user may want to ask “what time are you closing today?” or “how do I reset my router?”

By incorporating Search in the backend, Chat can answer questions with great accuracy, as it will pull from data that you already have stored in Yext.

Instruction Actions

Each Instruction has an action type that it will perform. You can create instructions for your bot with any of the following:

  • Collect: Ask the user for information.
  • Search: Look up information from Yext Content using Yext Search.
  • REST API: Send a request to a REST API (possibly using collected data).
  • Function: Execute a Yext serverless function (possibly using collected data).
  • Reply: Reply to the user. This is typically the last step in the instructions.

Logging

Yext will log the full history of each conversation users have with your chatbot so that you have complete visibility and can debug or optimize any goals you wish.

The Response Logs page will display every instance of a user message and the bot’s corresponding reply, along with the time, the predicted goal, and the reply status.

Integrating Chat

There are several different ways that you can integrate Yext Chat into your digital experiences.

We offer an out-of-the-box UI via a simple script tag. This option is an extremely easy and fast way to implement Yext Chat on any page. Using the Yext Chat CSS bundle, you can easily override key styling elements, such as fonts and colors.

Additionally, to ensure maximal flexibility and convenience for developers, Yext Chat provides three Javascript SDKs you can use to build applications for the web:

  • Chat Core
  • Chat Headless
  • Chat UI React

Additional Resources

unit Quiz
+20 points
Daily Quiz Streak Daily Quiz Streak: 0
Quiz Accuracy Streak Quiz Accuracy Streak: 0
    Error Success Question 1 of 2

    Which of the following are Instruction Types that you can create for your bot? (Select all that apply)

    Error Success Question 2 of 2

    Which of the following steps will take to create a Chat bot? (Select all that apply)

    Climbing that leaderboard! 📈

    You've already completed this quiz, so you can't earn more points.You completed this quiz in 1 attempt and earned 0 points! Feel free to review your answers and move on when you're ready.
1st attempt
0 incorrect
Feedback