Home / Beginner's Guide

How to Train Your Rule-Based Assistant

Anna Tsaturyan

Last updated on May 27, 2024

There are two types of assistants in Hoory AI. One is rule-based, the other is knowledge-based. To start training your AI Assistant, you must choose one of them.

If you prefer to train your AI Assistant manually, choose the knowledge-based assistant. Check out our user guide on training the knowledge-based assistant.

In this article, we'll walk you through training the Rule-Based Assistant.

Let's get started.

Rule-Based Assistant

The Rule-Based Assistant follows a structured approach, using dynamic flow charts to guide user interactions through pre-defined conversation flows.

The training options are as follows:

  • Train the assistant from scratch to customize its behavior entirely.
  • Use a pre-existing template to expedite the development process.
  • Import an unfinished file to continue working on a previously saved project.

The ready-made conversation templates include structures fit for

quizzes, product recommendations, NPS surveys, FAQs, lead scoring, and whatnot.

Training From Scratch

If you train your Rule-Based Assistant from scratch, you will be given ample space to create your custom conversation flow. The starting point will be given by default.

In the left-side panel, you can find many steps, fields, and functions to add to the flow. Here are the options:

Bubbles

  • Text
  • Video
  • Audio
  • Image
  • Embed

Inputs

  • Text
  • Email
  • Date
  • Button
  • Payment
  • File
  • Number
  • Website
  • Phone
  • Pic choice
  • Rating

Logic

  • Set variable
  • Redirect
  • Typebot
  • Jump
  • Condition
  • Script
  • Wait
  • AB Test

Integrations

  • Sheets
  • Analytics
  • Email
  • Make.com
  • Hoory AI
  • OpenAI
  • Webhook
  • Zapier
  • Pabbly
  • Pixel

Once you create the first draft of your conversation flow, you can preview it by clicking the "Preview" button in the top left corner. Then, if everything looks good, you can click Publish, and the flow will activate. Another option is to export the flow by clicking the three dots right under the Publish button and selecting "Export flow."

****