The Conversational AI block allows learners to explore your learning materials and helps check their understanding through an AI-powered chat.
Overview
The AI Conversation block appears as a chat between the AI and the learner. Depending on the selected mode, the AI will either:
- Tutor the learner on topics from your materials, or
- let the learners Practice by simulating a situation and asking the learner how they would respond.
Tutor mode
Tutor mode helps learners explore a topic through conversation while also checking their understanding. The AI introduces the key topics from your content and asks questions to help confirm whether the learner has understood the material. It teaches through scaffolding and interactive questioning while analyzing and working with a learner's level of understanding.
Practice mode
Practice mode lets learners practice by simulating a situation and prompting students to describe how they would respond. It engages learners in interactive roleplay situations where their choices and actions shape outcomes.
Availability
This feature is currently in beta. If you'd like to test it, please submit a request to our support team.
Note that you must have access to the Scenario Builder publishing in order to use this feature.
How to set up an AI Conversation block
- Create a new block and select 'Conversational AI'.
- Fill out the Title
- Fill out the Context with the materials the AI should use
- Choose Tutor mode or Practice mode
- Fill out the AI Prompt to customize how the AI will behave
- Optional - use the Customize tab to further customize the conversation
- Learning objectives let you define the objectives of the conversation
- Intro field lets you display a custom introductory message
- Situation field (Practice mode only) lets you describe the situation the AI should simulate and define the roles of the learner and the AI
- Preview the block to review how the AI functions
Title: this is the title of the block. It is shown to the learner, so use something short and descriptive.
Context: this is the text material the AI will use during the conversation. Include the topics, facts, and background you want the AI to cover. The AI will stick to the information you provide in this field, but it may use general knowledge if needed.
You can use the 'Copy from scenario' button to pre-fill the Context with other text content in your scenario, or you can use 'Extract from PDF' to extract text from one of your PDF files.
Mode: choose between Tutor mode and Practice mode described in the Overview section.
- Tutor mode is best for explaining topics and checking learners' understanding through various question formats.
- Practice mode is best for simulating different situations and applying practical knowledge.
AI Prompt: use this field to customize and control how the AI will behave. You can use this to define tone, level of detail, who your learners are, and include any other guidance.
Customize tab options:
- Learning objectives field lets you define the objectives of the conversation. Use this to narrow down what topics the AI should cover.
-
Intro field lets you display a custom introductory message that will be shown before the conversation begins. You can use this to provide brief instructions on how to interact with the AI.
If left empty, the AI will start the chat immediately after the learner reaches the block. -
Situation field (Practice mode only) lets you describe the situation the AI should simulate and define the roles of the learner and the AI.
If left empty, the AI will choose a situation and roles based on the context. We recommend filling this out to ensure a consistent experience for all learners.
Preview & test
After setting up the block, preview it to make sure it behaves as expected and covers the right topics.
Hover over the block in the editor and select Preview:
FAQ
What data does the AI Conversation block collect?
Currently, the AI Conversation block tracks only how much time each learner spends interacting with the block.
In the future, you will be able to review conversation outcomes (for example, whether the learner completed, skipped, or did not achieve the learning objective) and review the conversation itself.
Chat transcripts are not stored.
What languages are supported?
The AI Conversation block is currently available in English only. Support for additional languages is planned.
Are there usage limits?
There are no usage limits. Learners can have as many conversations with the AI as needed. The AI Conversation block currently does not contribute to your AI usage limits.
What AI model powers Conversational AI?
Conversational AI is powered by Google Gemini. The interactions via Conversational AI are not used to train or improve the model. Gemini does not receive any data about the user beyond the inputs provided via the chat interface, and no data is stored once the conversation is over.
See this article for more information on how your data is processed when using AI features on ThingLink.
Comments
0 comments
Please sign in to leave a comment.