Scenario Builder 2.0: AI tools

Scenario Builder 2.0 offers a number of AI tools - from simple features for working with text to a brand-new tool for converting your existing training manuals and materials into interactive scenarios.

Availability

AI tools within Scenario Builder 2.0 are currently available to all users. Note that this may change in the future.

If you have an organizational account, access to AI tools is governed by your organizational settings. By default, students cannot access any AI tools, including the tools described in this article.

Convert PDF training materials to scenarios

One of the core new features of Scenario Builder 2.0 is the ability to transform your existing training materials into interactive scenarios. Simply create a new scenario, click 'Create with AI', upload your existing training materials as a PDF file, and let our tools do their magic:

The AI will automatically determine the training topics and content areas based on the PDF's content. You can refine the content areas the AI should cover and choose between a linear and branched structure.

Once the AI processes the file, you will get a draft of your scenario including text blocks, multiple-choice questions that check the learner's knowledge as well as placeholder blocks that you can replace with interactive media content.

Enhance your text blocks

Convert text blocks to interactive scenes 

Let AI generate an interactive 2D image based on what you described in your text block. In addition to generating the image, the AI will reuse the text you provided to create hotspots.

Simply open a text block and click the 'Convert to ThingLink Scene' button:

Note: this action will replace the text block. If you do not want to lose the text inside the block, copy it first.

We highly recommend this option if you find that your scenario is too text-heavy.

Generate images within text blocks

Let AI generate an image that matches what you describe in your text block. The image will be added to the text block without replacing any text.

To use this tool, open a text block and click the blue 'Suggest matching image' icon:

Rewrite text blocks

Ask AI to rephrase the text used within the text block.

Generate new blocks based on the content of the scenario

Generate questions

You can use AI to generate 5 multiple-choice questions based on the content of your scenario or the text you provide. Question blocks contain4 options and usually 1 correct answer. You can always edit the question blocks if you would like to modify the questions or the answers available to your students.

Simply select the tool when adding a new block, then check what text it has selected and click the 'Create 5 questions' button. 

Suggest branches

The 'Suggest branches' tool generates branching paths based on the content of your scenario or the text you provide. You can then edit the paths or populate them with content

Select this tool when adding a new block, check the text it has selected, and click the 'Suggest branches' button.

Write a recap

The 'Write a recap' tool creates a text block with a summary of all preceding texts or the text you provide. Try using this tool after a series of text blocks or at the end of your branches to give your learners a quick summary of the content of each branch.

Simply select the tool when adding a new block, check the text it will summarize, and click 'Summarize'.

Limitations & troubleshooting

AI assistant tools for scenarios utilize only text from text blocks. Text from other block types (e.g. question blocks, the content of ThingLink scenes that are part of the scenario) and other types of content (e.g. images) are not utilized. When adding a new block using AI, the tool will take into account all texts that precede the block you are about to add. 

AI models & data flows

When transforming your PDFs, you can choose between three different AI models:

• Gemini 1.5 Flash is the fastest model we recommend for most content. This is the default model ThingLink preselects for you.
• Gemini 1.5 Pro is a slightly slower model. Flash and Pro produce similar results, but the latter works better if you have a complex PDF or if you are trying to transform multiple files at the same time.
• Azure GPT 4o is OpenAI's flagship model. Just like Gemini 1.5 Pro, it is slightly slower but better suited for working with complex materials.

Note that if you use Gemini 1.5 models, the PDFs will be processed using Google Cloud servers. If you use Azure GPT 4o, the data will be processed on Microsoft Azure servers controlled by ThingLink.

ThingLink uses Google's IMAGEN for generating images when you use the 'Suggest matching image' or 'Convert to ThingLink Scene' tools. IMAGEN runs on Google Cloud servers. 

Text-based tools ('Improve writing', 'Generate questions', 'Suggest branches', and 'Write a recap') use Azure OpenAI models run on servers controlled by ThingLink.

ThingLink does not use customer data to train AI models. ThingLink retains the content generated within the scenario but does not store the original PDFs or any additional data.